sampling &

sources of bias
census vs. sample

sources of bias

sampling methods

Dr. Mine etinkaya-Rundel

Duke University

census
Wouldnt it be better to just include everyone
and sample the entire population, i.e. conduct
a census?
Some individuals are hard to locate or measure,
and these people be different from the rest of the
population.
Populations rarely stand still.

Listen to the NPR story at http://www.npr.org/templates/story/story.php?storyId=125380052

exploratory
analysis
representative
sample

inference
Image credit: Wonderlane CC BY 2.0 http://www.flickr.com/photos/wonderlane/6231888661

a few sources of sampling bias

Convenience sample: Individuals who are easily accessible
are more likely to be included in the sample

Non-response: If only a (non-random) fraction of the
randomly sampled people respond to a survey such that the
sample is no longer representative of the population

Voluntary response: Occurs when the sample consists of
people who volunteer to respond because they have strong
opinions on the issue

Poll source: edition.cnn.com, August 29, 2013

1936

Landon vs. FDR

(Republican)

(Democrat)

Lose with 43% of the votes

Election results

Win with 62% of the votes

Image sources: http://en.wikipedia.org/wiki/File:LandonPortr.jpg, http://en.wikipedia.org/wiki/File:FDR_in_1933.jpg, and http://en.wikipedia.org/wiki/File:LiteraryDigest-19210219.jpg

Image:
http://www.flickr.com/photos/wonderlane/6231888661
Image credit: Wonderlane CC BY
2.0 http://www.flickr.com/photos/wonderlane/6231888661

sampling methods

simple random sample (SRS)

Stratum
Stratum 11
Stratum 1

Cluster
Cluster 99
Cluster 9

Cluster
55
Cluster
Index
Cluster
Index 5

Cluster
Cluster 33
Cluster 3

Cluster
Cluster 44
Cluster 4

Cluster
Cluster 77
Cluster 7

Index

Cluster
Cluster 11
Cluster 1

Stratum
Stratum 55
Stratum 5

cluster sample

Cluster
Cluster 22
Cluster 2

Stratum 6
Stratum 6
Stratum 6

Stratum
Stratum 33
Stratum 3

Stratum 4

Stratum 4
Index Stratum
4
Index
Index

Stratum 2
Stratum 2
Stratum 2

stratified sample

Cluster
Cluster 88
Cluster 8

Cluster
Cluster 66
Cluster 6

simple random sample (SRS)

Stratum 6
each case is equally likely to be selected
Stratum 2

Stratum 4

Index

Stratum 3

stratified sample

Stratum 4

Stratum 6

Index

Stratum 3

Stratum 1

Stratum 2

Stratum 5

Cluster
9
divide
the
population
into
homogenous
strata,
then
Cluster 5
Cluster 2
Cluster 7stratum
randomly sample from within each
Index

Cluster 3

Cluster 4

Cluster 8

cluster sample

Cluster 3

Cluster 4

Cluster 6

Cluster 8

Cluster 7

Cluster 9

Index

Cluster 5

Cluster 2

Stratum 5

Cluster 1

divide the population clusters, randomly sample a few

clusters, then randomly sample from within these clusters