Sampling distributions BPS chapter 11 © 2006 W. H. Freeman and Company

advertisement
Sampling distributions
BPS chapter 11
© 2006 W. H. Freeman and Company
Objectives (BPS chapter 11)
Sampling distributions

Parameter versus statistic

The law of large numbers

What is a sampling distribution?

The sampling distribution of

The central limit theorem

Statistical process control
x
Reminder:


Parameter versus statistic
Population: the entire group of
individuals in which we are
interested but can’t usually
assess directly.
A parameter is a number
describing a characteristic of
the population. Parameters
are usually unknown.

Sample: the part of the
population we actually examine
and for which we do have data.

A statistic is a number
describing a characteristic of a
sample. We often use a statistic
to estimate an unknown
population parameter.
Population
Sample
The law of large numbers
Law of large numbers: As the number of randomly-drawn observations
(n) in a sample increases,
the mean of the sample (x) gets
closer and closer to the population
mean m (quantitative variable).

the sample proportion (p̂
) gets
closer and closer to the population
proportion p (categorical variable).
What is a sampling distribution?
The sampling distribution of a statistic is the distribution of all
possible values taken by the statistic when all possible samples of a
fixed size n are taken from the population. It is a theoretical idea—we
do not actually build it.
The sampling distribution of a statistic is the probability distribution of
that statistic.
Note: When sampling randomly from a given population,

the law of large numbers describes what happens when the sample size n
is gradually increased.

The sampling distribution describes what happens when we take all
possible random samples of a fixed size n.
Sampling distribution of x
(the sample mean)
We take many random samples of a given size n from a population
with mean m and standard deviation s.
Some sample means will be above the population mean m and some
will be below, making up the sampling distribution.
Sampling distribution of “x bar”
Histogram
of some
sample
averages
For any population with mean m and standard deviation s:
The mean, or center of the sampling distribution of
population mean m.

x , is equal to the
The standard deviation of the sampling distribution is s/√n, where n
is the sample size.

Sampling distribution of
s/√n
m
x

Mean of a sampling distribution of
x:
There is no tendency for a sample mean to fall systematically above or
below m, even if the distribution of the raw data is skewed. Thus, the mean
of the sampling distribution of x is an unbiased estimate of the population
mean m —it will be “correct on average” in many samples.

Standard deviation of a sampling distribution of
x:
The standard deviation of the sampling distribution measures how much the
sample statistic
x
varies from sample to sample. It is smaller than the
standard deviation of the population by a factor of √n.  Averages are less
variable than individual observations.
For normally distributed populations
When a variable in a population is normally distributed, then the
sampling distribution of x for all possible samples of size n is also
normally distributed.
Sample means
If the population is
N(m,s), then the sample
means distribution is
N(m,s/√n).
Population
IQ scores: population vs. sample
In a large population of adults, the mean IQ is 112 with standard deviation 20.
Suppose 200 adults are randomly selected for a market research campaign.

The distribution of the sample mean IQ is
A) exactly normal, mean 112, standard deviation 20.
B) approximately normal, mean 112, standard deviation 20.
C) approximately normal, mean 112 , standard deviation 1.414.
D) approximately normal, mean 112, standard deviation 0.1.
C) approximately normal, mean 112, standard deviation 1.414.
Population distribution: N (m = 112; s = 20)
Sampling distribution for n = 200 is N (m = 112; s /√n = 1.414)
Application
Hypokalemia is diagnosed when blood potassium levels are low, below
3.5mEq/dl. Let’s assume that we know a patient whose measured potassium
levels vary daily according to a normal distribution N(m = 3.8, s = 0.2).
If only one measurement is made, what's the probability that this patient will be
misdiagnosed hypokalemic?
z
(x  m)
s
3.5  3.8

0.2
z = 1.5, P(z < 1.5) = 0.0668 ≈ 7%
If instead measurements are taken on four separate days, what is the
probability of such a misdiagnosis?
( x  m ) 3.5  3.8
z

s n
0.2 4
z = 3, P(z < 1.5) = 0.0013 ≈ 0.1%
Note:
Make sure to standardize (z) using the standard deviation for the sampling distribution.
Practical note


Large samples are not always attainable.

Sometimes the cost, difficulty, or preciousness of what is studied limits
drastically any possible sample size.

Blood samples/biopsies: no more than a handful of repetitions
acceptable. Often we even make do with just one.

Opinion polls have a limited sample size due to time and cost of
operation. During election times, though, sample sizes are increased
for better accuracy.
Not all variables are normally distributed.

Income is typically strongly skewed for example.

Is
x still a good estimator of m then?
The central limit theorem
Central Limit Theorem: When randomly sampling from any population
with mean m and standard deviation s, when n is large enough, the
sampling distribution of
x
is approximately normal: N(m,s/√n).
Population with
strongly skewed
distribution
Sampling
distribution of
x for n = 2
observations

Sampling
distribution of
x for n = 10
observations
Sampling
distribution of
x for n = 25
observations
Income distribution
Let’s consider the very large database of individual incomes from the Bureau of
Labor Statistics as our population. It is strongly right-skewed.

We take 1000 SRSs of 100 incomes, calculate the sample mean for
each, and make a histogram of these 1000 means.

We also take 1000 SRSs of 25 incomes, calculate the sample mean for
each, and make a histogram of these 1000 means.
Which histogram
corresponds to the
samples of size
100? 25?
How large a sample size?
It depends on the population distribution. More observations are
required if the population distribution is far from normal.

A sample size of 25 is generally enough to obtain a normal sampling
distribution from a strong skewness or even mild outliers.

A sample size of 40 will typically be good enough to overcome extreme
skewness and outliers.
In many cases, n = 25 isn’t a huge sample. Thus,
even for strange population distributions we can
assume a normal sampling distribution of the
mean, and work with it to solve problems.
Sampling distributions
Atlantic acorn sizes (in cm3)
14
– sample of 28 acorns:
12
Frequency
10
8
6
4


Describe the histogram.
What do you assume for the
population distribution?
2
0
1.5
3
4.5
6
7.5
Acorn sizes
What would be the shape of the sampling distribution of the mean:

for samples of size 5?

for samples of size 15?

for samples of size 50?
9
10.5 More
Further properties
The Central Limit Theorem is valid as long as we are sampling many
small random events, even if the events have different distributions (as
long as no one random event has an overwhelming influence).
Why is this cool?
It explains why so many variables are normally distributed.
Example: Height seems to be determined by a large number of
genetic and environmental factors, like nutrition.
So height is very much like our sample mean x.
The “individuals” are genes and environmental
factors. Your height is a mean.
Now we have a better idea of why 
the density
curve for height has this shape.
Statistical process control
Industrial processes tend to have normally distributed variability, in part
as a consequence of the central limit theorem applying to the sum of
many small influential factors. Random samples taken over time can
thus be used to easily verify that a given process is not getting out of
“control.”
What is statistical control?
A variable that continues to be described by the same distribution when
observed over time is said to be in statistical control, or simply in
control.
Process-monitoring
What are the required conditions?
We measure a quantitative variable x that has a normal distribution.
The process has been operating in control for a long period, so that we
know the process mean µ and the process standard deviation σ that
describe the distribution of x as long as the process remains in control.
An
x
control chart displays the average of samples of size n taken at
regular intervals from such a process. It is a way to monitor the process
and alert us when it has been disturbed so that it is now out of control.
This is a signal to find and correct the cause of the disturbance.
x
control charts
For a process with known mean µ standard deviation σ, we calculate
the mean
x
of samples of constant size n taken at regular intervals.
Plot x (vertical axis)
against time (horizontal axis).

Draw a horizontal center
line at µ.

Draw two horizontal
control limits at µ ± 3σ/√n
(UCL and LCL).

An
x
value that does not fall within the two control limits is evidence
that the process is out of control.
A machine tool cuts circular pieces. A sample of four pieces is
taken hourly, giving these average measurements (in 0.0001
inches from the specified diameter).
Because measurements are made from the specified diameter,
we have a given target µ = 0 for the process mean. The process
standard deviation σ = 0.31. What is going on?
x
xx x
x
x
For the
x
chart, the
center line is 0 and
the control limits are
±3σ/√4 = ± 0.465.
Sample
x
1
−0.14
2
0.09
3
0.17
4
0.08
5
−0.17
6
0.36
7
0.30
8
0.19
9
0.48
10
0.29
11
0.48
12
0.55
13
0.50
14
0.37
15
0.69
16
0.47
17
0.56
18
0.78
19
0.75
20
0.49
21
0.79
The process mean has drifted. Maybe the cutting blade is getting dull, or a
screw got a bit loose.
Download