Sampling
January 9, 2007
Cardinal Rule of Sampling
• Never sample on the dependent variable!
– Example: if you are interested in studying
factors that lead to organizational failure, you
cannot just study organizations that failed, you
need to compare them with organizations which
are successful
Types of Sampling
Probability Sampling
– Basic principle: a sample will be representative
of the population from which it is selected if all
members of the population have an equal
chance of being selected (EPSEM)
– Allows estimation of sample’s
representativeness
Types of Sampling
Probability Sampling
– Basic principle: a sample will be representative
of the population from which it is selected if all
members of the population have an equal
chance of being selected (EPSEM)
– Allows estimation of sample’s
representativeness
Nonprobability sampling
Sampling Concepts
• Element – unit about
which information is
collected
Sampling Concepts
• Element
• Universe – theoretical
aggregation of all
elements
Sampling Concepts
• Element
• Universe
• Population –
theoretically specified
aggregation of
elements
Sampling Concepts
• Element
• Universe
• Population
• Survey population –
aggregation of
elements from which
the survey sample is
selected
Sampling Concepts
• Element
• Universe
• Population
• Survey population
• Sampling unit –
element considered for
selection at some stage
of sampling
Sampling Concepts
• Element
• Universe
• Population
• Survey population
• Sampling unit
• Sampling frame – list
of sampling units from
which the sample is
selected
Sampling Concepts
• Element • Observation unit –
• Universe element from which
• Population data is collected
• Survey population
• Sampling unit
• Sampling frame
Sampling Concepts
• Element • Observation unit
• Universe • Variable – set of
• Population mutually exclusive
• Survey population characteristics
• Sampling unit
• Sampling frame
Sampling Concepts
• Element • Observation unit
• Universe • Variable
• Population • Parameter – summary
• Survey population description of a given
• Sampling unit variable in the
population
• Sampling frame
Sampling Concepts
• Element • Observation unit
• Universe • Variable
• Population • Parameter
• Survey population • Statistic – summary
• Sampling unit description of a given
• Sampling frame variable in a survey
sample
Sampling Concepts
• Element • Observation unit
• Universe • Variable
• Population • Parameter
• Survey population • Statistic
• Sampling unit • Sampling error – can
• Sampling frame be estimated using
probability theory
(standard error)
Sampling Concepts
• Element • Observation unit
• Universe • Variable
• Population • Parameter
• Survey population • Statistic
• Sampling unit • Sampling error
• Sampling frame • Confidence intervals –
computing sampling
error allows estimation
of confidence that the
parameter is within a
specified range
Types of Sampling Designs
• Simple random sampling
– Assign a number to each element in the
sampling frame then use a table of random
numbers to select elements for the sample
• Systematic sampling
– Every kth element in the total list is included in
the sample
– Sampling interval, periodicity
Types of Sampling Designs
• Stratified sampling
– Obtain greater representativeness and decrease
probable sampling error
– Organize the population into homogenous
subsets (with heterogeneity between subsets)
and select the appropriate number of elements
from each subset
Types of Sampling Designs
• Multistage Cluster Sampling (general or
stratified)
– Repeat 2 basic steps: listing and sampling
– List primary sampling units, (stratify if useful
to do so), take a sample, use sample to create
list of secondary sampling units (stratify), take
sample, etc.
– Each stage produces sampling error
Types of Sampling Design
• Probability Proportionate to Size (PPS)
– Each cluster is given a chance of selection
proportionate to its size
– The same number of elements is chosen from
each selected cluster
– This allows for selection of more clusters,
representation of elements in large clusters,
each population element has an equal chance of
selection
Types of Sampling Design
• Modifications to PPS designs
– May decide to include all very large clusters,
but then each element should be given the same
chance of being selected as in other clusters
– small clusters may contain fewer than the
number of elements to be taken per cluster, but
clusters can be combined to solve this problem
(try to combine homogenous clusters)
Nonprobability Sampling
• Purposive or judgmental sampling
• Quota sampling
• Reliance on available subjects
• Snowball sampling