week 8 sampling

Question 1

population vs sample

Answer 1

Population=complete set of subjects (usually very large). Sample=subset of the population. Usually use this to draw inferences re population. Inferences are only accurate if sample is truly representative of population.

Question 2

representative vs bias sample

Answer 2

representativeness=extent to which characteristics of the sample accurately reflect those of the population. To try to get most representative, try to reduce bias as much as possible. Bias may be due to chance or may be due to selection bias.

Question 3

Probability sampling methods

Answer 3

Meet the following conditions:

1. The entire population is known,
2. Each individual has a specifiable probability of selection.
3. Sampling is done by a random process.

Sometimes might think know an entire population (eg government list) but might still miss some.

Includes these methods:

simple random sampling

systematic sampling

stratified random sampling

proportionate stratified random sampling

cluster sampling.

Probability sampling has good chance of being representative, are time consuming, requires the researcher has in depth knowledge of the population.

Question 4

Non-probability sampling methods

Answer 4

Have the following conditions:

1. The population is not completely known
2. Individual probabilities of being selected cannot be known
3. Sampling is based on common sense and ease, with an effort to maintain representativeness and avoid bias.

Includes the following methods;

convenience sampling

quota sampling

purposeful sampling

Question 5

target population

Answer 5

The target population needs to be defined. Do we want to target whole population? is it ethical to?

Question 6

accessible population

Answer 6

Even if know every possible target in population, they may not be accessible. Therefore targt population further shrinks to accessible population.

Question 8

sampling error

Answer 8

when the sample mean or variance etc differs from those of the population.

3 causes; a) sampling bias (have selected for something)

b) chance. With larger samples, expect vagaries of chance to ‘even out”
c) Recording/measurement error (non sampling error) eg faulty equipment

sampling error occurs in random samples only. non sampling error occurs due either to deficiency or inappropriate analysis/measurement of data and may be in random or non random samples

Question 9

biased sample

Answer 9

occurs due to selection practices. eg. if considering gender roles in upper leadership and if only give survey to businesses striving to redress gender inequality, then our sample will contain a sampling bias.

Question 11

simple random sampling

Answer 11

Type of Probability Sampling.

Population clearly defined and each person in population identified. random sampling then used via method:

a) sampling with replacement. ie after being selected, subject is returned to “lottery”. possible to be selected multiple times. eg jury duty. or
b) sampling without replacement eg market research study. ie once selected, are excluded from possible re-selection.Technically then, samples are not quite independent and probability for selection changes eg in population of 1000, 1st draw is 1 in 1000, 2nd draw is 1 in999 etc.

Question 12

systematic sampling

Answer 12

type of Probability Sampling.

Same as per simple random sampling for the 1st participant, and after that, select every nth element.Not all participants can be selected. Might be biased, depending on how list comprised to give every nth participant etc.,

Question 13

stratified random sampling

Answer 13

Type of Probability Sampling

Population is first divided into strata (subgroups). Then select equal random samples from each strata as per simple random sampling.Most useful when researcher wants to compare subgroups. But SES etc may not reflect total population.

Question 14

proportionate stratified random sampling

Answer 14

Type of Probability Sampling

Begin by identifying a set of subgroups in the population.Determine what proportion of the population corresponds to each subgroup. Obtain samples from the subgroups with sample proportions of each subgroup exactly matching the population proportions.

Question 15

cluster sampling

Answer 15

Type of Probabilty Sampling.

All population is broken into groups or clusters. (eg school classes). then randomly select clusters of interest, and all subjects in a chosen cluster are sampled.Independence of participants is not guaranteed eg all students in a class share the same teacher, which may impact upon the research.

Question 16

convenience sampling

Answer 16

Type of Non-Probability Sampling

Also known as accidental sampling or haphazard sampling.

Participants selected either on availabiity or willingness.

Most common method in human research due to convenience.

Strong possibility of selection bias.need to provide clear descriptoion of how samples obtained and summary of participants’ SES etc.

Question 17

quota sampling

Answer 17

Type of Non-Probability sampling. Does attempt to make the sampling representative.

First define subgroups eg 5 different cultural groups, then establish number of individuals required to be sampled(quota) from each subgroup.

May not truly reflect population strata and population strata may not be fully known.

Question 18

purposeful sampling

Answer 18

Type of Non-Probability sampling.

Here, the goal is not to generalise from the sample to the population, therefore representativeness is not an issue.

Elements of convenience or quota sampling but on much smaller scale. Often for small case and qualitative research. Often targets very specific individuals. Might involve in depth interviews or individual case reports.

Question 19

Qualitative Sampling

Answer 19

Has a wide array of techniques and samples.

Usually involves very small samples and often in-depth interviews. Often done to reach a point of saturation (where new participants cannot offer any new insights). A sampling strategy is still required, based on theory.

Question 20

Sampling Summary

Answer 20

Selecting the right sample is crucial.

Probability samples are desirable for most research types but are harder to conduct.

Non -probability samples whilst less rigorous and generalisable, are still very valuable.