Samples and Bias

Simple Random Sample (SRS)

1. Number the "population"

2. Use a random way to select those numbered

3. Record the Data

Cluster Random Sample

1. Define what makes a cluster

2. Number the clusters and use a random way to select them

3. Indicate the entire cluster is used

Stratified Sample

1. Split the population into groups

2. use a method to random;y select from each group

3. Record the Data

Convenience Sample (Judgment Sample)

The person collecting the data chooses who to collect from (whatever is easiest)

Voluntary Response Sample

Participants choose whether or no to participate in the sample

Bias

When certain responses are systematically favored over others

Undercoverage Bias

part of the population has a reduced chance of being included in the sample

Question (response) Bias

Questions that are asked are confusing or leading

Nonresponse Bias

Data is not obtained from a particular part of the population (similar to undercoverage)

Disadvantage of Voluntary response sample

over represents people with strong opinions

Disadvantage of convenience sample

Opens the way for Bias, undercoverage bias, and question bias

Advantages of simple Random Sample

-free of bias

-cheap and easy for small populations/samples

-each sampling unit has a known and equal chance of selection

Disadvantages of simple random sample

-Takes Longer

-Lower chance of variety

-Sample may not represent population

Advantages of cluster random sample

Easy-Only select the number of cluster needed (usually low)

Disadvantages of cluster random sample

-May not represent the whole population

-Each member doesnt have an equal chance of being chosen

Advantages of stratified random sample

Guarentees representation from each strata (group)

Disadvantages of stratified random sample

-Difficult to do if you must divide the stratum

-Each member of population doesnt have an equal chance of being chosen

multi-stage sampling

Use a random method to select large groups and then repeat the method getting smaller groups from the previous groups each time. (Barrel-Box-Bag-Pouch-etc) (big -> small)

Systematic Sampling

select some starting point and then select every kth element in the population (use a magic number as range)

Response Bias

anything in a survey design that influences responses