Sample Statistics are used to estimate Population Parameters

Download Report

Transcript Sample Statistics are used to estimate Population Parameters

Sample Statistics are used to estimate
Population Parameters
We will use two different statistics:
For categorical data: the sample
proportion, p
For numeric data: the sample mean
Sample Statistics

Have sampling distributions
• Shape: normal if you follow the ‘rules’
• Center: equal to the parameter we’re
•
estimating if we take a random sample
Spread: related to the population standard
deviations by a factor of 1/n
Sample Proportion, p
• Shape: Normal if n and n(1-) ≥10
• Center: (p) = 
• Spread:  (p) =(1-)/n
Example for sample proportions:

Toss a coin 30 times. The probability of
getting a head is 80%.
• This is a binomial trial because:
• Each toss is independent of all of the other tosses
• There is a fixed number of tosses, n = 30
• There is a fixed probability of success,  = 0.80
Example con’t

What is the distribution of our sample proportion?
•
•
•
Shape: Normal if n and n(1-) ≥10
• n = 30*0.80 = 24,
• n(1-) = 30*0.20 = 6,
• so we can’t say the shape is normal
Center: (p) = 
• (p) =  = 0.80
Spread: (p) =(1-)/n
• (p) =(1-)/n = (0.8*0.2/30) = 0.073
Sample Mean,
• Shape:
Normal if
• the original data is normal, X~N(x, x2), or
• n is large, at least 30
• Center: (
• Spread: (
) = x
) =/n
Example of a sample mean,

If X~N(15, 22), what is the distribution of
• Since X is normal X 36 is also normal
• The mean is the same, 15
• The standard deviation is 2/36 = 1/3
• So, X 36~ N( 15, (1/3)2)
X 36 ?