Research: Population and Sampling - key terms, examples, advantages and disadvantages

Convenience Sampling

A non-probability sample based on using people who are easily accessible - such as in shopping malls. One of the most common sampling methods. Example: You just ask someone who walks past you to answer your questions. Advantages: Easiest, saves time, money and effort. Disadvantages: Poorest rationale for selection, lowest credibility, may be biased, may not be representative of whole population. Disadvantage example 1: Researcher may only choose people because they look more helpful or friendly or because they are located closest. Disadvantage example 2: The person selected may not meet the research criteria (eg. too old/young, wrong job, etc.)

Snowball Sampling

A non-probability sample in which additional respondents are selected based on referrals from initial respondents. Refer a friend. "Can you recommend someone else to participate..."

Quota Sampling

A non-probability sample in which quotas (total numbers/percentages) are based on chosen demographic factors. The researcher chooses, so it is not random. Example 1: 20% each in the study should be Native American, African American, Mexican American, Asian American, and European American. Example 2: 15 1st year students, 15 2nd year, 15 3rd year.

Purposive (Judgement) Sampling

A non-probability sample in which the selection criteria are based on the researcher's personal opinion about who should be in the study. The researcher samples with a particular purpose in mind, and chooses subjects because of certain characteristics and knowledge of population.

Random Sample

A probability sample in which everyone has an equal chance of being selected. A powerful sampling method Example: Can use a random number generator in Excel to choose subjects from a list. Example: 100 random students chosen from all Engineering courses at HCT.

Systematic Sample

A probability sample in which the entire population is numbered and elements are selected using a numbered interval. Example 1: a shop carrying out customer loyalty research chose every 10th and 15th customer on their list and kept going down. Example 2: Every 7th student on the list of Engineering students at DMC. Advantages: Easier than random sampling. Better coverage of study area than random sampling. Disadvantages: More biased than random - not all members have equal chance of being selected.

Stratified Sample

A probability sample where researcher divides the population into groups based on characteristics, and then randomly selects from each group. Example: Comparing the level of diabetes symptoms between upper socioeconomic and lower socioeconomic status.

Sampling Error

Error that occurs because the sample selected is not perfectly representative of the population. Example sampling error: In a company with 80% males and 20% females, the researcher selected 10% of the males, and 90% of the females. This is not representative.

Sampling (definition and examples)

In a sample, data are gathered on a small part of the whole population and are used as an example of what the whole population picture is like. Example: the total possible population for the research is 500 students, the researcher chooses a sample of 50 students to survey. Sampling examples: 1. Probability sampling (random, stratified and systematic), and 2. Non-probability sampling(convenience, purposive/judgement, quota and snowball).

Non-probability Sampling

Non-probability sampling is used when a complete population list is NOT available. Non-probability sampling is commonly used in qualitative research. It is not random. Advantages: 1. Useful for generating ideas and obtaining feedback - eg. qualitative research. 2. Convenient. 3.Lower cost. 4. Easier than probability sampling. Disadvantages: 1.Cannot generalize results to entire population with high level of confidence. 2.Less reliable than probability sampling (harder to repeat the research and get similar results). 3. More biased than probability. Non-probability sampling examples: Convenience, Purposive (Judgment), Quota and Snowball.

Probability Sampling

Probability sampling is used for quantitative research. Probability sampling can be used when you have contact information for a known whole population, and can select a random sample from your population. Example 1: You have a list of ALL students in HCT Engineering courses - this is your whole population. You can then select your sample based on a mathematical calculation of probability. Example 2: You have a fair idea of the size of a large population - eg. The population of Dubai was estimated at just over 2 million in 2013. Probability examples: Random sampling, Systematic sampling, Stratified sampling. Advantages: 1.You can generalize results from a random, systematic or stratified samples to represent the entire population. 2.More reliable / powerful than non-probability sampling. 3.Avoid bias. Disadvantages: 1. More expensive, 2. more time-consuming, 3.mre complex than non-probability. (Convenience, purposive, quota and snowball sampling are cheaper, easier and faster)

Sample size - quantitative

The sample size is larger for quantitative research (eg. surveys/questionnaires) and smaller for qualitative (eg. interviews, focus groups, observation). You can use statistical analysis (eg. probability methods such as random and systematic sampling) to help you choose a representative sample for quantitative research. Because the sample size is larger in quantitative, it is more likely that you can generalize your results to the whole population.

Sample size - comparison of quantitative and qualitative

The sample size is usually much smaller for qualitative research (eg. interviews, focus groups, observation) and larger for quantitative research (eg. surveys or questionnaires). For example, you might want to interview ten people, or have eight people in your focus group to obtain your in-depth qualitative data. In comparison, you might survey 30 or 50 or 500 people to get your numerical/quantitative data.

Population

Typically we think of a population in terms of the number of people in a country (eg. approx. 4.5 million people in New Zealand in 2014). A research population is made up of the people or cases we want to study. Examples of research populations: - Students enrolled at HCT - Students studying Business at DMC - Users of Facebook or Twitter

Sampling - why do we sample populations?

We take a sample of the population because of logistics and practical issues. There is not enough time, energy, money, labour, equipment, or available access to measure every single item within the main population. Example: As an individual researcher, you cannot survey ALL the residents of New Zealand, or ALL Facebook users, or you cannot interview ALL students at HCT, or ALL students in Engineering. You want to select a portion or sample that represents the whole population.

Sample size - qualitative

You often use non-probability sampling (eg. Convenience, Purposive/Judgment, Quota or Snowball) to help you choose a sample for qualitative research. Qualitative sample size is based on the idea of data "saturation" - you keep going until you get enough detailed information, or the same ideas are starting to be repeated. Because the sample size is smaller in qualitative, it is NOT likely that you can generalize your results to the whole population.