In this example, we have a worksheet containing the names of all of the Department of Statistics' full-time faculty members from the Spring 2021 semester.
These data are in the following files. The file ending in .mwx is a Minitab worksheet file; this can only be opened with Minitab 20. The file ending in .xlsx is an Excel file; this can be opened with any version of Minitab as well as with Excel:
FacultySP21.mwx
FacultySP21.xlsx
If this is your first time opening an .mwx file you may receive an error message if your computer does not know to open this in Minitab. You should be able to fix this by saving the file to your desktop, opening Minitab, and then opening the worksheet from within Minitab. After the first time, you computer should recognize that .mwx files should be opened with Minitab.
To select a simple random sample of 10 names from this dataset, follow the steps below. At the bottom of this section there is a video that shows where to click.
- Open the data in Minitab
- From the tool bar, select Calc > Sample from Columns...
- In the Number of rows to sample box, enter 10
- Click in the From columns box and then double click the Name variable
- Click in the Store samples in box and type MySample
- Click OK
The third column of your worksheet should now be labeled "MySample" and it should contain 10 names. Since we are using simple random sampling procedures, the results will be different each time due to random sampling variation. Try these steps a few times, you should see that you get a different set of 10 names each time.
Video Walkthrough
When we hear the term population, the first thing that comes to mind is a large group of people.
In market research, however, a population is an entire group that you want to draw conclusions about and possesses a standard parameter that is consistent throughout the group.
It’s important to note that a population doesn’t always refer to people, it can mean anything you want to study: objects, organizations, animals, chemicals and so on.
For example, all the countries in the world are an example of a population — or even the number of males in the UK. The size of the population can vary according to the target entities in question and the scope of the research.
When do you need to collect data from a population?
You use populations when your research calls for or requires you to collect data from every member of the population. Note: it’s normally far easier to collect data from whole populations when they’re small and accessible.
For larger and more diverse populations, on the other hand — e.g. a regional study on people living in Europe — while you would get findings representative of the entire population [as they’re all included in the study], it would take a considerable amount of time.
It’s in these instances that you use sampling. It allows you to make more precise inferences about the population as a whole, and streamline your research project. They’re typically used when population sizes are too large to include all possible members or inferences.
Let’s talk about samples.
What is a sample?
In statistical methods, a sample consists of a smaller group of entities, which are taken from the entire population. This creates a subset group that is easier to manage and has the characteristics of the larger population.
This smaller subset is then surveyed to gain information and data. The sample should reflect the population as a whole, without any bias towards a specific attribute or characteristic. In this way, researchers can ensure their results are representative and statistically significant.
To remove unconscious selection bias, a researcher may choose to randomize the selection of the sample.
Types of samples
There are two categories of sampling generally used – probability sampling and non-probability sampling:
- Probability sampling, also known as random sampling, is a kind of sample selection where randomization is used instead of deliberate choice.
- Non-probability sampling techniques involve the researcher deliberately picking items or individuals for the sample based on their research goals or knowledge
These two sampling techniques have several methods:
Probability sampling types include:
- Simple random sampling
Every element in the population has an equal chance of being selected as part of the sample. Find out more about simple random sampling. - Systematic sampling
Also known as systematic clustering, in this method, random selection only applies to the first item chosen. A rule then applies so that every nth item or person after that is picked. Find out more about systematic sampling. - Stratified random sampling
Sampling uses random selection within predefined groups. Find out more about stratified random sampling. - Cluster sampling
Groups rather than individual units of the target population are selected at random.
Non-probability sampling types include:
- Convenience sampling
People or elements in a sample are selected based on their availability. - Quota sampling
The sample is formed according to certain groups or criteria. - Purposive sampling
Also known as judgmental sampling. The sample is formed by the researcher consciously choosing entities, based on the survey goals. - Snowball sampling
Also known as referral sampling. The sample is formed by sample participants recruiting connections.
Find out more about sampling methods with our ultimate guide to sampling methods and best practices
Calculating sample size
Worried about sample sizes? You can also use our sample size calculator to determine how many responses you need to be confident in your data.
When to use sampling
As mentioned, sampling is useful for dealing with population data that is too large to process as a whole or is inaccessible. Sampling also helps to keep costs down and reduce time to insight.
Advantages of using sampling to collect data
- Provide researchers with a representative view of the population through the sample subset.
- The researcher has flexibility and control over what kind of sample they want to make, depending on their needs and the goals of the research.
- Reduces the volume of data, helping to save time.
- With proper methods, researchers can achieve a higher level of accuracy
- Researchers can get detailed information on a population with a smaller amount of resource
- Significantly cheaper than other methods
- Allows for deeper study of some aspects of data — rather than asking 15 questions to every individual, it’s better to use 50 questions on a representative sample
Disadvantages of using sampling to collect data
- Researcher bias can affect the quality and accuracy of results
- Sampling studies require well-trained experts
- Even with good survey design, there’s no way to eliminate sampling errors entirely
- People in the sample may refuse to respond
- Probability sampling methods can be less representative in favor of random allocation.
- Improper selection of sampling techniques can affect the entire process negatively
How can you use sampling in business?
Depending on the nature of your study and the conclusions you wish to draw, you’ll have to select an appropriate sampling method as mentioned above. That said, here are a few examples of how you can use sampling techniques in business.
Creating a new product
If you’re looking to create a new product line, you may want to do panel interviews or surveys with a representative sample for the new market. By showing your product or concept to a sample that represents your target audience [population], you ensure that the feedback you receive is more reflective of how that customer segment will feel.
Average employee performance
If you wanted to understand the average employee performance for a specific group, you could use a random sample from a team or department [population]. As every person in the department has a chance of being selected, you’ll have a truly random — yet representative sample. From the data collected, you can make inferences about the team/department’s average performance.
Store feedback
Let’s say you want to collect feedback from customers who are shopping or have just finished shopping at your store. To do this, you could use convenience sampling. It’s fast, affordable and done at a point of convenience. You can use this to get a quick gauge of how people feel about your store’s shopping experience — but it won’t represent the true views of all your customers.
Manage your population and sample data easily
Whatever the sample size of your target audience, there are several things to consider:
- How can you save time in conducting the research?
- How do you analyze and compare all the responses?
- How can you track and chase non-respondents easily?
- How can you translate the data into a usable presentation format?
- How can you share this easily?
These questions can make the task of supporting internal teams and management difficult.
This is where the Qualtrics CoreXM technology solution can help you progress through research with ease.
It includes:
- Advanced AI and machine learning tools to easily analyze data from open-text responses and data, giving you actionable insights at scale.
- Intuitive drag-and-drop survey building with powerful logic, 100+ question types, and pre-built survey templates. For more information on how to get started on your survey creation, visit our complete guide on creating a survey.
- Stylish, accessible and easy-to-understand reporting that automatically updates in real time, so everyone in your organization has the latest insights at their fingertips.
- Powerful automation to get up and running quickly with out-of-the-box workflows, including guided setup and proactive recommendations to help you connect with other teams and react fast to changes.
Also, the Qualtrics online research panels and samples help you to:
- Choose a target audience and get access to a representative sample
- Boost the accuracy of your research with a sample methodology that’s 47% more consistent than standard sampling methods
- Get dedicated support at every stage, from launching your survey to reporting on the results.