Sampling Part 1

1
SamplingPart 1

• November 3, 2004
• Sociology 487

2
Lecture Topics

• Sampling methods
• Nonprobability
• Probability
• Sample components
• Logic of Probability Sampling

3
Observation and Sampling

• Polls and other forms of social research, rest on
  observations
• The task of researchers is to select the key
  aspects to observe or sample
• When is sampling unnecessary?
• All units of the population are identical
• All units of the population can be studied
• U.S. Census
• In all other circumstances, necessary to sample
  members of our population

4
Sampling Components

• Population entire set of individuals / other
  entities to which study findings are to be
  generalized
• Sample subset of the population which is used to
  study the population as a whole
• Elements individual members of the population
  whose characteristics are measured
• Sampling frame a list of all elements or other
  units containing the elements in a population

5
Sampling Methods

• Probability sampling
• Involves random selection of sample elements
• Yields a representative sample
• A sample that looks like the population from
  which it was selected
• Nonprobability sampling
• Does not involve random selection of sample
  elements
• Sample may be unrepresentative
• Some characteristics will be over-represented and
  some characteristics will be under-represented in
  comparison with the prevalence in population

6
Generalizability

• Sample generalizability
• Do our sample results reflect reality in study
  population?
• Cross-population generalizability
• Do our sample results reflect reality in target
  population?
• Target population a set of elements larger than
  or different from the population that was sampled
  and to which researchers would like to generalize
  any study findings

7
Nonprobability Sampling

• Sampling methods that do not let us know in
  advance the likelihood of selecting each element
• General characteristics
• Often used in qualitative research
• Used in quantitative research when probability
  selection is not possible
• Useful in preliminary, exploratory studies
• Does not yield representative samples
• Reduces ability to generalize findings to larger
  population

8
Availability / Convenience Sampling

• Selection of elements is done by what is
  available and/or convenient
• Does not result in a representative sample
• Useful in exploratory studies
• Often used in popular research

9
Quota Sampling

• Quota pre-set number of elements based on
  characteristics in a population to ensure that
  sample represents those characteristics in
  proportion to their prevalence in the population
• Problems
• To determine quotas, must know relevant
  characteristics of entire population
• Sample representative in terms of quota
  characteristics but probably not representative
  in other ways

10
Quota Samples, continued
11
Purposive Sampling

• Selecting a sample on the basis of knowledge of
  a population, its elements, and the purpose of
  the study
• Selecting informants
• Knowledge of topic of study
• Willingness to talk
• Representativeness of range of viewpoints
• Continue to select informants until you have
• Completeness
• Saturation

12
Snowball Sampling

• Collect data on members of population who can be
  located then ask those individuals to help locate
  other members of that population.
• When is snowball sampling appropriate
• members of a population are difficult to locate
• exploring populations of interest before
  developing formal sampling plan
• Problems
• No confidence about how representative sample is
• Initial contacts may shape entire sample

13
Respondent-Driven Snowball Sampling
14
Recap Non Probability Sampling Methods

• Do not know in advance the probability of
  selecting sample elements
• Study samples are not representative
• Difficult to determine whether results can be
  generalized to larger population

15
Probability Sampling

• Used when researchers want precise, statistical
  descriptions of large populations
• A sample of individuals from a population must
  contain the same variations that exist in the
  population for results to be generalizable

16
Political Polls and Survey Sampling

• One of the most visible uses of survey sampling
  is political polling that is then tested by
  election results.
• In the 2000 Presidential election, pollsters came
  within a couple of percentage points of
  estimating the votes of 100 million people.
• To gather this information, they interviewed
  fewer than 2,000 people.

17
Election Eve Polls - Voting for U.S.Presidential
Candidates, 2000
18
Things to Consider When Evaluating Research Based
on Probability Sampling

• Is probability of selection random?
• How is sampling frame defined?
• What is the response rate?
• Sampling error
• Larger the sample, smaller the sampling error

19
Election Eve Polls - Voting for U.S.Presidential
Candidates, 2004
20
Political Polls

• Population Probable voters in the United States
• Registered to vote, Intention to vote, Voting
  history, Interest in Presidential campaign, Age,
  Knowledge of polling place, Voting for the 1st
  time
• Sample in Gallup Poll, 1,573
• Elements individual likely voters
• Sampling Frame List of all telephone numbers in
  the United States