Probability Sampling

Front 1

Why do we use a sample?

Back 1

To draw conclusions from about a larger group of people or respondents.

Front 2

Why is it rare that every possible respondent in a population is included in a study?

Back 2

Because of time and costs

Front 3

What does practical limitations mean?

Back 3

Information about a subset (sample) of the population must be treated as representative of the entire population.

Front 4

What are inferential statistics used to make?

Back 4

Generalizations about a larger group of people from data gathered from a smaller subset of the larger group.

Front 5

Is it likely that results four in a sample are identical to those obtained if the entire population was measured?

Back 5

Nope

Front 6

What is a sampling frame?

Back 6

The target group under study

Front 7

What is the representation of the sampling frame?

Back 7

Sample

Front 8

What is the sample size?

Back 8

Number of respondents in the study

Front 9

What is a sampling method?

Back 9

How respondents are selected

Front 10

What is the response rate?

Back 10

The percentage of sample who provide valid responses

Front 11

Describe sampling frame

Back 11

Sampling frame is a listing that includes all the elements or units in a survey population. The sampling frame is unlikely to be perfectly accurate so the goal becomes establishing the most representative sample possible.

Front 12

Describe comprehensiveness of sampling frame

Back 12

A sample can only be representative of the sampling frame.

Front 13

What is the efficiency of the sampling frame?

Back 13

If the sampling frame excludes ppl from pop. sample estimates will be biased. eg. only contacting ppl with phone leaves out ppl without phone. If sampling frame includes ppl from pop that aren't part of study, sample with be biased. eg. people doing study for reward who don't meet sampling frame characteristics

Front 14

What is and is not necessary about probability of selection from a sampling frame?

Back 14

1. It is necessary that the probability of being selected for a study is known.

2. It is not necessary that every individual has the same probability of being selected for study

Front 15

What is a representative sample?

Back 15

A sample drawn in such a way that its characteristics exist in about the same proportions as in the pop. from which it was drawn.

Front 16

How accurately a sample reflects the pop. depends on what?

Back 16

1. sample frame
2. sample size
3. design of selection procedure

Front 17

How must a sample be representative?

Back 17

Selection of the sample must be random. By chance, the sample will be more representative than if we asked for volunteers, chose ppl or selected haphazardly.

Front 18

What do the laws of probability do to sampling error?

Back 18

It does not eliminate sampling error but it does reduce it. An estimate made from a sampling frame is unlikely to be perfectly related to the pop.

Front 19

What 3 sources of error/bias exist in sampling?

Back 19

1. If a non-probability or non-random sampling method is used
2. If the sampling frame is inadequate
3. If some members of the pop. refuse to participate or cannot be contacted (non-response bias) those who participate in a study may systematically differ from those who do not.

Front 20

What is the margin of error?

Back 20

Techniques that exist to calculate a reasonable CI for an estimate made from a sample. Sampling errors and biases increase the likelihood of drawing incorrect conclusions. The margin of error is directly related to sample size.

Front 21

What does the required sample size depend on?

Back 21

1. Heterogeneity of the population

2. Degree of accuracy required in our conclusions

Front 22

What is heterogeneity of the population?

Back 22

If the population is homogeneous, only a few observations are needed to generalize. The more hetero, the larger the sample needed.

Front 23

What determines the degree of accuracy in a sample size?

Back 23

The larger the sample the closer it is to the population. As the sample size increases the variability of the statistic decreases.

Front 24

What are the two methods of sample selection that exist?

Back 24

1. Probability sampling

3. Non-probability sampling

Front 25

What are probability sampling techniques?

Back 25

Each unit in a sample has a known chance for inclusion in the sample. eg. simple random sampling, systematic sampling, stratified sampling, multi-sage cluster sampling

Front 26

What is the best method for obtaining a representative sample?

Back 26

Random selection where every member of the pop. from which a sample is drawn has an equal chance of being selected. Selection of one member has no effect on the selection of any other member.

Front 27

What are the two types of random sampling?

Back 27

1. Sampling with replacement

2. Sampling without replacement

Front 28

What is simple random sampling?

Back 28

Sampling with and without replacement.

Front 29

What is systematic random sampling?

Back 29

Every Kth element in the sampling frame is chosen for study. More efficient than random sampling. Starting point in from must be randomly selected. Periodicity or bias due to repetition or cyclical quality in sampling frame may occur.

Front 30

What is stratified sampling?

Back 30

Ensures samples contain enough causes of a particular sub pop and or guarantee equal numbers of pop. characteristics

Front 31

What is multi-stage cluster sampling?

Back 31

With large pops a better sampling strategy might be to target specific areas and draw samples from only those delimited spaces. Eg. cities, then schools, then classes then students.

Front 32

What's an advantage of multi-stage cluster sampling?

Back 32

The greater the heterogeneity of the clster and the finer the stratification (smaller the clusters involved) the more precise the results.

Front 33

What's a disadvantage of multi-stage cluster sampling?

Back 33

Clusters tend to be internally homogeneous

Front 34

What are non-probability sampling techniques?

Back 34

Do not provide potential respondents with a known chance for inclusion in the sample. Does not involve random selection whereas probability sampling does.

Front 35

Is non-probability sample generalization to the larger pop.?

Back 35

It's difficult to know whether the sample represents the larger population.