Topic 12: Research Methodology: Sampling

Question 1

What are the stages in the selection of a sample?

Answer 1

1. Define the target population
2. Determine the sampling frame
3. Determine whether probability or non-probability sampling method is appropriate(appropriate sampling technique)
4. Determine sample size
5. Execute sample process

Question 2

Questions to ask when defining the population for sampling?

Answer 2

• Is there a list available that matches our population? If so, can it be used? Is there valid contact information and can they be reached by an appropriate communication method?
• Who are we NOT interested in?
• What are our relevant market segment characteristics
• Do we only want a regional population? If so, how do we determine the borders?
• Should the study include multiple populations?

Question 3

What is a sampling frame and what is sampling frame error?

Answer 3

• A sampling frame is a list of population elements from which a sample can be drawn. aka a working population
• Sampling frame error is when a sampling frame doesn’t include all the elements of a population, or does not accurately represent the target population

Question 4

What is the difference between probability and non-probability sampling?

Answer 4

• Probability sampling involves a sample where each member of the target population has a known, non-zero probability of being chosen
• Non-probability sampling is when each member of the target population has an unknown probability of being chosen
• Although probability sampling is preferred, sometimes non-probability sampling is more practical, and if done correctly, can introduce minimal bias/error.

Question 5

What are the different types of non-probability sampling techniques?

Answer 5

• Convenience sampling- Choosing respondents that are most conveniently available; quick, easy, cheap, but cannot represent a population other than that sample (Good for exploratory research)
• Judgement sampling- Experienced researcher chooses sample based on their own judgement of what qualities they are looking for
• Quota sampling- ensuring all various subgroups of a population are represented to the extent that the researcher desires by assigning quotas to each subgroup (Advantages over probability sampling: speed of data collection, lower costs and convenience)
• Snowball sampling- intial respondents selected using probability techniques, but subsequent respondents selected by those initial respondents (useful when target population is narrow/difficult to reach) Can introduce bias since subsequent respondents are likely to be similar to initial respondents

Question 6

What are the different types of probability sampling techniques?

Answer 6

• Simple random sampling- using random number generator to determine which units of a population are included in the sample
• Systematic sampling- random starting point selected, and then every nth number of the population is chosen for the sample (divide population by sample size to determine systematic interval)
• Stratified sampling-dividing population into strata(subgroups) based on some common quality within that stratum, and then using simple random selection within those strata (can be proportional or disproportional)
• Cluster sampling (individual element is replaced with a cluster of elements in close proximity; clusters are randomly chosen)
• Multistage area sampling

Question 7

What is the difference between proportionate and disproportionate stratified probability sampling?

Answer 7

• Proportionate stratified sampling is when the number of sampling units drawn from the stratum is proportional to the population size of that stratum
• Disproportionate stratified sampling is when the sample size for each stratum is determined by analytical considerations (reasons: as stratum variability increases, sample size must also increase to provide more accurate data; reduce random sampling error)

Question 8

What factors should be considered when deciding on an appropriate sample design?

Answer 8

• Degree of accuracy necessary- exploratory research often doesn’t need samples to be as accurate/representative as causal research
• Resources- some sampling techniques are more expensive than others; consider firm’s resources available for the project
• Time- projects with deadlines might benefit more from simpler, quicker sampling methods
• Advance knowledge of the population (are there sufficient lists of the target population available?
• National or local project- geographic proximity of population will influence sample design