While working on a Six Sigma project, at some point the team members will be required to choose a sample size of the individuals being studied. For this aim, a number of sample techniques are covered during Lean Six Sigma training courses. During the Six Sigma phase called “Measure” from the Six Sigma “Define, Measure, Analyze, Improve, Control” (DMAIC) cycle, a Six Sigma sample is typically obtained.
The two primary categories of sampling techniques—probability samples and non-probability samples—will be distinguished throughout a Six Sigma Green Belt course. Let’s take a deeper look at the 5 types of sampling methods — both categories as well as their subcategories.
What Is A Probability Sampling Method?
The choosing of elements from a population is done using probability sampling techniques, which take into account known probabilities. Simple random sampling, probability related to number of respondents, etc. are a few examples. When using probability sampling techniques, it is feasible to ascertain both the likelihood that every sample is going to be chosen and which units correspond to each sample. Basically, this technique selects the sample with a specified probability.
Probability sampling techniques include the following examples:
- Simple Random Sampling
- Cluster Sampling
- Stratified Sampling
- Systematic Sampling
Sampling Method: Simple Random Sampling
When discussing probability sampling techniques, simple random sampling (SRS) is a particular instance for random samples. If every unit included in the population can be matched with an equal probability of being chosen inside the sample, then that sample is said to be called an SRS. SRS is a fundamental sampling approach where we choose a subset of people for research from a broader population. Every person is randomly selected, and each person has a match or equal probability of being selected for the sample. Keep in mind that the probability of selection is the same for any feasible sample of any size. Another name for it can be unrestricted random sampling (URS).
SRS techniques come in two different varieties. Simple random sampling without replacement is the first, while simple random sampling with replacement is the second. Sampling w/replacement occurs when an element from the population can be chosen more than once. Sampling with no replacement occurs if an element of the population only has the chance of being chosen once.
Sampling Method: Cluster Sampling
Spreading your sample throughout the entire population might occasionally be costly. We could select cluster sampling techniques while evaluating sample strategies with regard to cost. The population is first divided into clusters or groups using cluster sampling. Randomly selected clusters are chosen to reflect the population. The sample is then expanded to cover every unit contained within the chosen clusters. The population is then represented by the chosen clusters.
There are no components inside the sample representing clusters that were not chosen. They are portrayed by individuals from specific clusters. In contrast to stratified sampling, which chooses some units out of each group, this is different. Industries, schools, and geographical locations like electoral subdivisions may be a few types of clusters.
The sample techniques of cluster sampling, as well as stratified sampling, seem to be significantly dissimilar. Using cluster sampling, you may categorize the population (clusters), obtain data for each sampling unit that’s located on every randomly chosen cluster, and take an SRS of a certain number of clusters from the available clusters.
How Does Cluster Sampling Work?
It is simpler to describe how various sampling techniques differ from one another with the earlier example. According to the same situation that was used to describe stratified sampling, we’re using a population that includes all first-graders in the nearby school district. The strata in the following row include twenty distinct primary schools. By using SRS, six elementary schools were chosen from the twenty primary schools. Additionally, every kid in the six elementary schools that were chosen is included inside of the sample size.
Particulars |
Example |
Population |
All primary students in the local school district |
Group (Strata) |
Twenty different primary schools in the local school district |
Simple Random Sample |
Six primary schools from the twenty possible |
Sample |
Every student in the six selected primary schools |
Sampling Method: Stratified Random Sampling
Among all sampling techniques, “stratified sampling” is the process that is most frequently employed in surveys. This method is mostly used to decrease population variety and to improve the accuracy of the estimations.
To stratify is to divide into groups. This approach separates the population into a series of strata or subgroups. Every stratum should be as homogenous or comparable as feasible when the strata are produced. The appropriate sample is then formed by taking a random sample from every stratum and combining them. Please note that a stratum refers to a single group.
Using stratified sampling you should: Divide all populations into smaller groups (strata), get an SRS from every group (stratum), and collect the data from each sampling that was sampled randomly from every group, for example, stratum.
Non-proportional and proportional stratified sampling techniques are also available. Subgroups (strata) are given equivalent and proportional representation for proportional sampling methods. The sample is going to have a larger size if there are more items and the other way around. ‘N’ stands for population size, while ‘n’ stands for sample size. Each stratum receives a portion of the sample size so that what constitutes the sample fraction can remain constant for every stratum. That is determined by n/N=c. Each stratum is thus represented in this way based on its size. Regardless of prevalence in the population or not, all of the sub-strata are equally represented when it comes to non-proportionate samples.
Stratified sampling has various advantages over basic random samples when it comes to sampling techniques, including the potential for lower survey costs per observation and greater precision at a given cost.
How Does Stratified Random Sampling Work?
Let’s take a look at stratified sampling by using the table below. Focus on the “particulars” row. The population’s range is discussed in the first column. The population includes all first-graders in the nearby school district. The strata include twenty distinct elementary schools. From each elementary school out of those twenty, fifty kids were chosen using basic random sample methods. Last but not least, one thousand pupils from twenty elementary schools make up this sample size.
Particulars |
Example |
Population |
All primary students in the local school district |
Group (Strata) |
Twenty different primary schools in the local school district |
Simple Random Sample |
Fifty students from each of the twenty possible schools |
Sample |
50 * 20 = 1000 selected students |
Sampling Method: Systematic Sampling
When it comes to systematic sampling, everything about the entire selection process is based on what can be considered a random beginning. With the use of random values, the 1st unit will be chosen, and the next units are chosen automatically in accordance with a pre-established pattern.
Using SRS, the starting spot and one of the 1st K elements will be chosen at random, and every K’th item in the framework is chosen for the sampling. As a result of its simplicity and convenience, this strategy is frequently used. Whenever a full breakdown of the entire population becomes available, systematic sampling is a regularly employed sampling technique. It can also be known as interval sampling or quasi-random sampling.
Please be aware that this technique is frequently employed in industry, when a product is chosen for evaluation from a manufacturing line, say, every 15 minutes, to make sure that the equipment and machinery are operating as intended. This method may also be applied when interviewing participants in a survey study. After choosing an individual at random to use as the beginning point, a research analyst can choose to question every fifteenth customer that enters a specific business, or after choosing a home arbitrarily as the starting point, they might interview the residents of every seventh house on a specific street.
In that scenario, it’s also possible that a researcher would like to use a defined size sample. Therefore in that instance, it is important to first familiarize oneself with the entire population size that the sample will be drawn from.
How Does Systematic Sampling Work?
Under Systematic Sampling, the Kth item will be chosen out of the frame from the sample, and can be referred to as the “sampling interval”. So, that would be the case if we wish to choose a sample number of fifty students out of five hundred students. Since N is equal to n/n, if we sample interval “K”, it would be the same. N stands for the population size, while lowercase n stands for the sample size. The letter K is therefore equivalent to ten, which can be obtained by dividing five hundred by fifty. The sample interval will be K = ten.
Selecting a number at random, such as “i”, “K” and all of the “Kth” units constitutes systematic sampling. Assuming that “i” is a random number, we choose five, fifteen, twenty-five, thirty-five, forty-five, etcetera. The “random start” is the random number “i”. “K” samples will be produced using the procedure with equal chance. When information needs to be gathered from plants inside a forest, buildings in blocks, serially ordered entries, and more, systematic sampling will be preferred over other sampling techniques.
Non-probability – The Second Main Sampling Method
Non-probability sampling techniques make up the second primary sampling strategy. Non-probability occurs when a sample gets deemed to be “representative” of a certain population or when judgment is used in order to choose “representative” units inside that population. Purposive sampling, as well as judgment, are the two names for this technique. Surveys of public opinion frequently employ this technique.
A quota sample is a typical judgment sample that is used in surveys. Due to the enumerator’s possible prejudice and/or bias, this approach is not generally employed. However, in the case that the enumerator is knowledgeable and skilled, this approach could produce useful findings. For instance, all new automobile buyers formed the sample for the market research study on the functionality of a newly released car.
Please take notice that non-probability sampling techniques should most likely be avoided wherever feasible. These sampling techniques rely on human decision-making instead of random selection. When it comes to such sampling techniques, possible sources of prejudice and/or bias are uncontrolled, hence, a statistical theory will be unable to describe how they could behave. When feasible, always employ probability sampling techniques.
Have you led projects and are looking to earn a project management certification? You might be interested in learning about how lucrative this can be. Check out these articles.
13 PMP Benefits Once You Get The PMP Certification
No experience leading projects but still want to get into project management? No problem! Check out these articles.
CAPM Certification Eligibility
What is a Certified Project Manager; How do I get PM Certifications