Thursday, March 3, 2011

Sampling: Design and Analysis – Sharon Lohr – Chapter 1 – Simple Probability Samples


 

-          Known probability of selection

 

Types of Probability Samples


-          Simple random sample

o   A SRS of size n is taken when every possible sub set of n units in the population has same chance of being the sample

-          Stratified random sample

-          Cluster sample

Framework for Probability Sampling


-          Sampling distribution

 

Simple Random Sampling


-          Simple random sample with replacement

o   In finite population sampling, sampling with replacement provides no additional information.

-          Simple random sample without replacement

-          Finite population correction

o   Make this correction because with small populations, the greater our sampling fraction, n/N, the more information we have about the population and thus smaller variance.

-          For large populations it is the size of the sample taken, not the percentage of population sampled that determines precision of estimator.

o   If soup well stirred, you need taste only 1/2 spoonfuls, whether you have made 1 or 20 litres.

 

Confidence Intervals

 

 

Sample Size Estimation

 

-          Specify tolerable error

-          Find an equation

-          Estimate unknown quantities

 

Systematic Sampling

 

-          Is technically a form of cluster sampling

-          Randomly chose first point, then take kth point

 

Randomization Theory Results for Simple Random Sampling

 

 

A Model for Simple Random sampling

 

-          Joint probability distribution … supplies link between units in the sample and units not in the sample.

 

 

When Should a Simple Random sample Be used

 

 

No comments: