Class Notes (1,100,000)
CA (620,000)
UTM (20,000)
SOC (4,000)
SOC211H5 (100)
Lecture 5

SOC211H5 Lecture Notes - Lecture 5: Sampling Error, Systematic Sampling, Simple Random Sample


Department
Sociology
Course Code
SOC211H5
Professor
kivanchi
Lecture
5

This preview shows half of the first page. to view the full 2 pages of the document.
Sampling error and non- response error
Sampling error: the extent to which characteristics of the sample deviate from the
characteristics of the population
It consists of:
1) Random sampling error: due to chance, assessable by probability calculus: results
are valid at .05 probability level, results are accurate 19 time out of 20
2) Systematic sampling error: due to sampling design, i.e. to the unequal probability
of sampling elements to be selected into the sample. Unequal probability results in
systematic differences between the selected and the unselected sampling elements
Non- response error occurs after the sample has been selected serious, because
participants nearly always differ systematically from participants who could not be found
or have refused to participate
The rich; the homeless, Aboriginals; over-representation of the elderly in survey
samples
To deal with (sampling and) non-response error: weigh (weight) the statistics by
demographic characteristics of the population
Population: 49% male, sample: 40% male: weigh each response by a man by 1.225
- 40% men/49% population
Types of Sampling Design
Probability sampling: probability of selection of each sampling elements is known (not
necessarily equal). Random sampling error is calculable.
It requires:
1) a sampling frame
Known-probability selection of sampling elements
Non- probability sampling: probability of selection of each sampling element is unknown,
therefore it must be presumed to by unequal. Random sampling error is incalculable
AND there is systematic sampling error.
Types of probability sampling
-Simple random sample: select sampling units from a sampling frame, using random
numbers
- Advantage: known and equal probability of selection (p=n/N).
You're Reading a Preview

Unlock to view full version