Qm - Module 6 - Sampling Methods

Front 1

Define population

Back 1

the complete set of all conceivable observations of a variable

Front 2

Define sample

Back 2

a subset of a population is a sample

Front 3

Why is information almost always collected from a sample rather than the population?

Back 3

1. Economic - less expensive
2. Timeliness - takes less time
3. Size and accessibility - difficult to collect from population
4. Observation and destruction - destruction of element being observed. Cannot destroy population.

Front 4

Describe the two theories associated with sampling

Back 4

1. How sample is chosen
2. How inferences are made about population from the sample.

Front 5

List several applications of Sampling

Back 5

1. Opinion polls
2. Quality control
3. Checking invoices (auditing)

Front 6

What must exist for correct conclusions to be drawn about a population?

Back 6

Samples must be representative.

Front 7

Define sampling bias

Back 7

Samples are consistently not representative of the population.

Front 8

What is the fundamental method of ensuring that samples are representative of the population?

Back 8

Using simple random sampling

Front 9

List and describe variations of simple random sampling.

Back 9

1. Multi-stage sampling - permits collection in just a few sections of area, cutting down on travelling and interviewing expenses.
2. Stratified sampling - using knowledge about population to determine how to sample.

Front 10

Define Judgement Sampling

Back 10

all methods that are not random and use personal judgement.

Front 11

Primarily process for random sampling

Back 11

using random generated numbers that are assigned to the population.

Front 12

Describe Multi-Stage Sampling

Back 12

1. Population is split into groups, each group split into sub-groups, etc.
2. Samples taken at each stage
Benefit is that it is easier to get listing of populatio in this manner.

Front 13

Describe Cluster Sampling

Back 13

Same as Multi-Stage Sampling except at the final grouping all individuals are included in sample.

Front 14

Describe Ordinary Cluster Sampling

Front 15

Describe Stratified Sampling

Back 15

1. Use prior knowledge of population to make sample more representative
2. Population is split into subpopulations (strata) of known proportion
3. Sample is taken of each strata in same proportion.

Front 16

How does weighting affect sampling

Back 16

If strata exists but proportions are not known before sampling, then weighting providing the adjustment after sampling. The strata's result are weighted according to this formula (% of pop/% of sample)

Front 17

Describe probability sampling

Back 17

A sampling method which allows for elements to have differing chances of being selected as a sample.

Front 18

Describe variable sampling

Back 18

A sampling method where some special subpopulation is over-sampled. Due to the importance of gathering information from this subpopulation.

Front 19

Describe Area Sampling

Back 19

Artificial breaking down of the population to make sampling easier.

Front 20

List the main Judgement Sampling methods.

Back 20

1. Systematic Sampling - taken at regular intervals from Population
2. Convenience Sampling - easiest way possible, when no other way exists. Medical. Must understand - not random.
3. Quota Sampling - to overcome interviewer bias. List given to interviewer about how many in each sub-group (random from there on).

Front 21

What is the problem of inference in sampling?

Back 21

Since no sample can be definitely known to be representative of the population, there is a need to estimate the errors that may arise. Can only be done when Random sampling used.

Front 22

How does stratified sampling affect accuracy of estimates?

Back 22

Increases level of estimates

Front 23

Define the sampling frame

Back 23

the complete list from which the sample is selected (not population - due to inaccurate records, etc.) A big difference between population and sampling frame may mean results are unreliable.

Front 24

Describe non-response and its impact

Back 24

If cannot complete the interview for, then that is non-response. This may lead to bias in the sample.

Front 25

List types of sampling bias

Back 25

1. Sampling frame error
2. Non-response
3. Inaccurate measurement
4. Interviewer bias
5. Interviewee bias
6. Instrument bias

Front 26

What are the approaches to determining Sample size?

Back 26

1. Ask what level of accuracy is needed? Can determine size through mathematics.
2. Collect largest sample that budget allows.

Front 27

Methods of sampling for small population

Back 27

1. Replacement - return element to population after selection (may be selected again)
2. Sampling without replacement - uses a different theory for calculating accuracy. More complicated.