In business research, primary data are collected from people or non-human objects such as countries or states.  Most typically, people provide the data for business research. When it's time to gather data, the researcher has to decide to collect primary data from the population or from a sample of the population of interest. Gathering data from a population would be time consuming, costly, and unmanageable for the business researcher.  Some argue that it is impossible to collect data from a population. Reasons might include that people aren't always where they belong, they move, they can't respond, they don't want to respond, and so forth. Common sense dictates that to avoid errors in collecting data, a sampling of a subgroup of a population would be less time consuming, less costly, and easier to manage. Again, some would point out that a valid sample is more representative. By sampling, the researcher will be able to control time and cost, control for errors, and generalize the results to the target population. 

Population – A group of potential participants to whom you want to generalize the results of a study. The total group of subjects or participants must have at least one common characteristic.

Subjects - People or objects involved in a research study.

Sample – A subset of the population. It is desirable that the sample be representative of the population.

Sampling Frame - A list of people or objects from the population.

Sampling Errors - Factors contaminating or influencing the results of who participates in the study. For example, a sampling error might occur if an important subgroup in the population is excluded.

Probability Sampling - Everyone in the target population has an equal and known chance of being selected for the study. Probability sampling designs utilize randomization for candidate selection.

Nonprobability Sampling – The probability of selecting any one member from the population is not known.

Sampling Plan - Researcher's plan of action for selecting the subjects or participants.

The Sampling Process:

Before selecting participants for a study, business researchers should take the following steps:

Select a sampling frame - A list of the population. Also, the list of potential participants in a study.

Select a sample design - The method of selecting study participants.

Select the appropriate size - This is the number (N) of subjects invited to be in the study. The N should be large enough for a 95% confidence level.

Select the sample - The researcher proceeds to select the participants. NOTE: Since not everyone who is invited to participate in a study will, researchers often over sample, assume a certain participation rate (e.g., 10%), and work until the final sample of participants is representative of the population.

How Large Does the Sample Need to Be?:

This is one of the most frequent questions asked of researchers. Today, there are a variety of methods for determining sample size. Some research textbooks contain a table for determining sample size from a given population. Other books, such as the Berenson et.al. text, provide very detailed and complex formulas for determining correct sample size. Below are some proportions cited in Krejcie, R.V. and Morgan, D.W. for determining sample size. (N = population size; S = sample size)

You should notice that the larger the population, the smaller the sample size in proportion to the population. Generally, the results of the study will be more valid with larger samples.

Today, there are many online tools for calculating sample size. Take a look at this site:

http://www.researchinfo.com/docs/calculators/samplesize.cfm

The concepts of sampling are very important. Managers need to know enough to question and evaluate sampling or other research procedures utilized in a research study. Managers need to know enough to ask experts "intelligent" questions. Since cost is a major factor of conducting research, sampling efficiency should always be considered.

Probability Sampling Designs:

Ideally, researchers would like to control possible biases in their studies, specifically in their samples.  By controlling biases, researchers enhance the internal and external validity of their findings.  One of the critical tools in controlling threats to validity (biases) is to conduct probability sampling using random selection of subjects.

Simple Random Sampling

Simple random sampling is a technique where we select a group of subjects (a sample) for the study from a larger group (a population). Each individual is chosen entirely by chance and each member of the population has an equal chance of being included in the sample. In the appendix of most research method books, a random numbers chart is available for random selection of subjects. Today, many computers are equipped with random numbers generators for sample selection. A sample that is the correct size and is randomly selected will be representative of its population.

Cluster Random Sampling

Formal groups of subjects, rather than individuals, are identified and randomly selected for the study. Cluster examples: (a) voting districts, (b) departments in an office building, and (c) classrooms in a school.

Systematic Random Sampling

Researchers use this sampling technique when their sampling frame already exists in some order (e.g., alphabetical). To minimize bias, they start at a random point in the list then select every "kth" name on the list. For example, if your sampling frame consisted of 1000 names in alpha order, and you want 100 people in your study, you can start your selection at a random point and continue selecting every 10th name until you have your 100 subjects.

Stratified Random Sampling

There may often be factors that divide up the population into important subgroups or strata. For example, consider a population of undergraduate students at a large university. They consist of first, second, third, and forth year students; males and females; handicapped students; athletes; older students; etc. Sometimes researchers believe that a subgroup might be especially hard to get into a sample (e.g., they are hard of hearing and don't respond to a knock on the door from an interviewer). We can ensure these important subgroups are not left out of the study by conducting stratified sampling. A stratified sample is obtained by randomly selecting people from each stratum or sub-group of a population. When we sample a population with several strata, we generally require that the numbers of individuals in each stratum be in proportion to their percentage in the population. For example, if 10% of a university population is handicapped, and you want to make sure handicapped students are represented in your sample in proportion to the population, your strata or subgroup of handicapped students should be 10% of your final sample.    

Multistage Random Sampling

A multistage random sample is constructed by taking a series of simple random samples in stages. This type of sampling is often more practical than simple random sampling for studies requiring "on location" analysis, such as door-to-door surveys. In a multistage random sample, a large area, such as a country, is first divided into smaller regions (such as states), and a random sample of these regions is collected. In the second stage, a random sample of smaller areas (such as counties) is taken from within each of the regions chosen in the first stage. Then, in the third stage, a random sample of even smaller areas (such as neighborhoods) is taken from within each of the areas chosen in the second stage. If these areas are sufficiently small for the purposes of the study, then the researcher might stop at the third stage. If not, he or she may continue to sample from the areas chosen in the third stage, etc., until appropriately small areas have been chosen.

Sampling and Statistical Inference

You have just read about the term random sampling. The use of randomization in sampling allows for the analysis of results using statistical inference. Statistical inference is based on the laws of probability, and allows analysts to infer conclusions about a given population based on results observed through random (representative) sampling.

Nonprobability Sampling Design:

In the real world, there are times when randomization is impossible.  Reasons include time, expense, and importance of the findings. It is likely you will not be able to conduct probability/random sampling for your project 2 because of time and money limitations. This is OK. When researchers cannot or do not want to conduct random sampling, they will select a nonprobability sampling design.  That is, the people in the population will not have an equal chance of being selected.  They will have an unknown probability of being in the study. Simply stated, this means the final sample may have sampling errors, it will not be representative of the population, and the results might not be valid and reliable. Next are examples of nonprobability sampling designs.

Judgment Sampling

The researcher chooses people or objects that might represent the purpose of the study.  For the manager, this design is less time consuming and less costly for the corporation. Exploratory research incorporates this design.

Convenience Sampling

This design uses the philosophy of inviting into a study anyone who is available. The researcher specifies the rules of selection. To earn extra class credits, college students agree to participate in their professor's research study.  The advantages of the design are less cost and less time consuming. Exploratory research and descriptive research may utilize convenience sampling. Many of you will use convenience sampling in your Project 2.

Snowball Sampling

This is one of my favorite designs. This type of sampling is conducted when a list of the population of interest does not exist. Consider research that is conducted on alcoholics or battered women. There is typically no accurate listing of a population of people with these characteristics. However, these are important groups in our society, and we often want to have them in our research projects. Therefore, we find one person with the characteristic of interest (e.g., alcoholism), we invite them into the study, then ask if they can suggest other individuals with this same characteristic that might participate in our study.

Next Steps:

• Complete all readings and assignments for Module 3.
• Want to learn even more about sampling beyond the lecture and course texts? Check out the syllabus "Other Useful Online Resources." The Trochim electronic textbook can be very useful.