Ch 11 Sampling
- Biases in sampling due to
o Not all students available at the same time, may have exclusions
o The decisions about whom to approach may be influenced by judgments.
o Anyone not enrolled in that specific course will get excluded.
- Element/ unit- a single case in the population, usually a person.
- Population-all cases about which one seeks knowledge. To which the researchers
conclusions are meant to apply.
- Sampling frame: the list of elements from which the sample will be selected.
- Sample- the elements selected for investigation, a subset of population. – may
involve probability or non probability
- Representative sample- a sample that is microcosm of the population.
Represents its essential characteristics. Used in probability sampling process.
- Probability sample- a sample selected using a random process. Gives everyone
equal chance to be picked. Minimizes sampling error.
- Non-probability sampling- a sample selected using non random method. Giving
some units a better chance to be selected.
- Sampling error- errors of estimation that occur if there is a difference between
the characteristics of sample and a population from which it was selected. Can
occur even in random sampling.
- Non-response- refusal of a unit to participate in a study, can’t be contacted, or
don’t supply the required data.
- Census: data collected from all elements in a population.
o The enumeration of all members of a population of a nation state- national
- Sampling ratio: n/N (sample size /total population)
- Sampling interval: every 20 participant.
- Three sources of bias:
o Not using a random method to pick the sample. Human judgment will
affect the selection process. Avoided by probability sampling.
o The sampling frame or list of potential subjects is inadequate- if excludes
cases, even random sampling cant help
o Some people in the sample refuse to participate or cannot be contacted-
non response, sample may be non representative.
- Sampling error
o Important to be aware of sampling error, ex: one too many in on category
o Important to know that sampling error can’t be completely eliminated.
o Can be reduced by probability sampling by the known probability of error
- Types of probability sample:
o Simple random sample- most basic form. Each unit has equal probability
• Define the population
• Select or devise a comprehensive sampling frame.
• Decide sample size
• List all students in population and assign them random
• Using a table of random numbers select n.
• The students that match the n number get included.
o No opportunity for bias & no issues with participant
• The use of table of random numbers- • Simple random sampling without replacement- not using a
same participant twice.
o Systematic sample:
Selected directly from sampling frame, without using random
Periodicity- no inherent ordering or pattern in the sampling frame
o Stratified random sampling
Stratifying the population/ dividing it in subgroups, by criterion
and selecting a simple random sample or systematic sample from
each of the resulting strata.
Ensures that the sample is distributed in the same way as in the
population in terms of stratifying criterion.
Is feasible only when it is relatively easy to identify and allocate
units to strata.
More then one stratifying criterion can be used.
o Multi staged cluster sampling
The primary sampling unit is aggregate of units of population.
Involves cluster and subunits within those clusters.
More economic in large populations.
No adequate sampling frame required.
Allows interviews to be far more geographically concentrated.
- Qualities of probability sample
o Allows one to make inferences from the sample to the population from
which it was selected.
o Can be generalized.
- Sample size:
o Effected by cost and time
o Absolute and relative sample size
Absolute size is important, not the proportion of the population it
Increasing the size of sample increases the precision of the
estimates derived from it.
Large sample can’t guarantee precision.
Large samples decreasing sampling error
o Non- Response
Refusal of some people, from participating in study.
Response rate: Percentage of the sample that participate in the
Response rate= Number of usable questionnaires x 100
Total sample- unsuitable or uncontactable
Members of the sample
there is a decline in response rate
Things like: subject matter of research, type of respondent and
level of effort expended on improving the number of cooperating
respondents affect response rates.
Mailed questionnaires: prone to non response
Following up with people who don’t respond is a good way to get
o Heterogeneity of the population
When population heterogeneous, samples drawn are likely to be
highly varied. Should have larger sample size to maximize
accurate representation. When homogeneous, variation is less.
o Kind of analysis:
Contingency table: showing relationship between two variables.
In a 2x2 table- here are four cells in which the cases can fall.
- Types of non-probability sampling
o Includes all forms of sampling that’s not conducted according to
probability sampling methods.
o Quota sample- considered as good as probability sampling.
Convenience sampling- when elements are readily available.
• Good for pilot studies.
• May not be used to test newly create scales for reliability or
to generate ideas for further research.
Snowball sampling- a form of convenience sampling
• Contact one small group of people who are relevant to the
research topic, and then use them to establish contacts with
• It’s not random.
• Very unlikely to be representative of the population.
• Used in qualitative studies.
• Best used when researcher wants to focus on the
relationships between people and tracing connections.
• Infrequently used in academic social research.
• Used in commercial research.
• Aim to produce a sample that reflects a population in terms
of the relative proportion of people in different categories.
• Is not carried out r