Stratified sampling is a sampling method used by researchers to divide a bigger population into subgroups or strata, which can then be further used to draw samples using a random sampling method. The stratified sampling technique is useful in ensuring that every subgroup, or stratum, within the population is adequately represented in the sample. Stratified sampling, due to its nature, offers several advantages over simple random sampling, such as increasing the precision and reliability of the results especially when there are significant variations between the strata.
The stratified random sampling method has widespread uses and is particularly useful in diverse populations where certain segments might otherwise be underrepresented. In this article, we will cover the basics of stratified sampling and answer key questions, including:
- What is stratified sampling?
- What are the uses of stratified sampling?
- What are the types of stratified random sampling?
- When should you use stratified random sampling in your research?
By understanding these aspects, researchers can effectively use stratified sampling to enhance the accuracy and representativeness of their studies.
What is stratified sampling?
Stratified sampling is a probability sampling technique in which the population is first divided into distinct, non-overlapping strata based on a specific characteristic, such as age, income level, or education. These strata should be mutually exclusive, meaning each group is internally homogeneous but different from other groups in terms of the characteristics being studied. No member should belong to more than one group. After identifying the strata, a random sample is drawn from each group. This process is known as stratified random sampling. The size of these samples can be proportional to the size of the strata (proportional stratified sampling) or equal across all strata (equal allocation).
Stratified sampling is preferred when the goal is to study distinct characteristics among groups and the population is diverse. It helps avoid the overrepresentation or underrepresentation of certain population features, which can occur with simple random sampling. By ensuring adequate representation of all strata, stratified random sampling reduces sampling bias, allowing for more accurate generalization of results. This method is invaluable in fields such as market research, public health studies, and educational research, where understanding differences among specific groups is crucial.
What is stratified random sampling?
In stratified random sampling, a larger population is divided into distinct subgroups, or strata, that share similar characteristics to study their appreciable differences. By dividing the population into strata that share similar characteristics, researchers can draw random samples from each group in proportion to their sizes in the overall population. This ensures that every subgroup is adequately represented, enhancing the accuracy and reliability of the results.
Here are the key steps in stratified random sampling:
- Identify the Strata: The first step is to divide the population into non-overlapping groups based on specific characteristics, such as age, gender, income level, etc. These groups are called strata.
- Determine the Sample Size: Decide how many individuals or items will be sampled from each stratum. This can be proportionate (where the sample size from each stratum is proportional to its size in the population) or equal (where each stratum is sampled equally regardless of its size).
- Random Sampling Within Strata: Conduct random sampling within each stratum. This ensures that every member of a stratum has an equal chance of being included in the sample.
- Combine the Samples: Combine the samples from all strata to form the final stratified random sample.
When to use stratified sampling?
Stratified sampling is ideal for studying differences between subgroups, especially in large populations. It is the method of choice when high accuracy, efficiency, and representation are essential. Here are situations where stratified sampling is preferred:
- Heterogeneous population: When the population is diverse and consists of distinct subgroups (strata) with different characteristics, stratified sampling can ensure adequate representation of each subgroup.
Example: In a population of patients with various medical conditions, stratifying by condition ensures that each condition is proportionately represented in the sample.
- Comparative studies: When making comparisons of characterisitics between different subgroups within a population, stratified sampling can be useful as it allows for more accurate and reliable comparisons between these groups.
Example: When comparing the effectiveness of a new treatment method for relieving migraine pain across different age groups, stratifying by age ensures that each age group is adequately represented in the sample.
- Ensuring representativeness: The stratified sampling method can ensure adequate representation of the groups that are smaller in number and hence might be underrepresented with simple random sampling. On the contrary, groups that are bigger in number pose a risk of overrepresentation with simple random sampling. In such cases, a stratfied sampling method can offer a solution by dividing the population into different groups from which samples can be drawn in a manner to reflect their proportion in the population.
Example: In a survey of a city’s population, if a particular ethnic group is a minority, stratifying by ethnicity ensures this group is included in the sample.
- Improving precision: Stratified sampling can reduce the overall variability by reducing the standard error in the estimate of the sample, leading to more precise and accurate results. Stratified sampling ensures low variation within the subgroup (with no or minimum overlap between the subgroups).
Example: In agricultural research, stratifying a sample by soil type can lead to more precise estimates of crop yield since soil type can significantly affect yield.
- Limited resources: Stratified sampling can be used when resources such as time and money are limited. It ensures that all relevant subgroups are studied without the need to increase the overall sample size excessively.
Example: In market research, if a company wants to understand the preferences of different customer segments, stratifying by age group or income level ensures efficient use of resources while obtaining detailed insights.
- Reducing sampling bias: Stratified sampling limits sampling bias, especially in populations where certain subgroups are more difficult to reach or are less likely to respond.
Example: In a political poll, stratifying by demographic factors like age, gender, and education level helps in obtaining a sample that more accurately reflects the voting population.
How to use stratified sampling?
Once you select stratified sampling as your method, follow these systematic steps to ensure your sample accurately represents the subgroups of the population. Here is a detailed guide on how to use stratified sampling:
- Define the Population: The first step in stratified sampling is to define the target population for your research. The complete and exhaustive list of all members of the population should be compiled for this purpose.
- Identify the Strata: Next, divide the population into distinct subgroups, or strata, based on specific characteristics that are relevant to your study. Common characteristics used for stratification include age, gender, income level, education, geographic location, etc.
- Determine the Size of Each Stratum: Calculate the size of each stratum. This involves counting the number of members in each subgroup.
- Decide the Total Sample Size: Decide on the total number of observations you want in your sample; this is basically the number of units you will have in total drawn from each group. This is decided based on the desired level of precision and the resources available.
- Allocate the Sample Size Proportionally: Allocate the total sample size to each stratum based on its proportion in the population. This ensures that the sample is representative of the entire population. Use this formula for proportional allocation:
Sample size of the stratum=(Sample size÷Population size)×Stratum size𝑆𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑡𝑟𝑎𝑡𝑢𝑚=(𝑆𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒÷𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑠𝑖𝑧𝑒)×𝑆𝑡𝑟𝑎𝑡𝑢𝑚 𝑠𝑖𝑧𝑒
- Random Sampling Within Each Stratum: Once the sample size is calculated, draw samples from each stratum by using random sampling to select the required number of observations. This ensures that every member of each stratum has an equal chance of being included in the sample.
- Combine the Samples: Draw and combine the samples from all strata to form the final stratified random sample.
- Perform Data Analysis: Since the sample is representative of the population, carefully analyze the data to get results that will provide accurate insights into the population as a whole and also allow for meaningful comparisons between strata.
Types of stratified random sampling
There are two common types of stratified random sampling—proportional and disproportional stratified sampling. The preferred type depends on how the samples are allocated and the goals of the study. Here are the types and their uses:
- Proportional Stratified Sampling: In proportional stratified sampling, the sample size for each stratum is proportional to the stratum’s size relative to the entire population. This ensures that each stratum is represented in the sample according to its proportion in the population.
Example: If a population consists of 60% females and 40% males, and you want to sample 100 individuals, you would select 60 females and 40 males so that more females are selected compared to males to reflect that actual occurrence in the population.
- Disproportional Stratified Sampling: In disproportional stratified sampling, the size of each stratum is not proportional to its size in the population. Considering the above example, the researcher may decide to select 30 females and 35 males, the goal of disproportionate sampling is to study underrepresented groups which might have characteristics that you are interested in studying.
Example: If you are studying three departments in a company with 200, 350, and 400 employees respectively, and you decide to sample 100 employees from each department, you are using disproportional stratified sampling.
Examples of Stratified Random Sampling
Here are some examples of how stratified random sampling can be a valuable tool across various disciplines:
Example 1: Health Research
Scenario: A public health researcher wants to study the effectiveness of a new drug against diabetes in a city with a population of 100,000 people. The population might have different responses among different age groups. The stratified random sampling can be applied to address the research question.
| Age group | 0-19 | 20-39 | 40-59 | 60 & above | Total |
| Number of diabetic people in strata (stratum size) | 10,000 | 30,000 | 40,000 | 20,000 | 100,000 (Population size) |
| Sample size of the stratum | 500 | 1,500 | 2,000 | 1,000 | 5,000 (Sample size) |
Proportional Stratified Sampling:
If the researcher decides to sample 5,000 people. The sample size for each stratum will be proportional to the population size of the stratum and can be calculated by using the formula:
Sample size of the stratum=(Sample size÷Population size)×Stratum size𝑆𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑡𝑟𝑎𝑡𝑢𝑚=(𝑆𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒÷𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑠𝑖𝑧𝑒)×𝑆𝑡𝑟𝑎𝑡𝑢𝑚 𝑠𝑖𝑧𝑒
Implementation:
The researcher will then randomly select 500 people from the 0-19 age group, 1,500 from the 20-39 age group, and so on, ensuring that each age group is adequately represented in the sample.
Example 2: Educational Research
Scenario: An educational researcher wants to evaluate the academic performance of 700 students in a school district with five high schools after receiving a new course on meditation. Each school has a different number of students who received the training. The stratification can be done accordingly:
| Schools | A | B | C | D | E | Total |
| Number of students receiving the training | 500 | 300 | 400 | 200 | 600 | 2,000 (Population size) |
| Sample size of the stratum | 175 | 105 | 140 | 70 | 210 | 700 (Sample size) |
Implementation:
The researcher will then randomly select 175 students from school A, 105 students from school B, and so on with a total of 700 students out of five schools and analyze the data.
Example 3: Market Research
Scenario: A company wants to conduct a survey on 15,000 customer satisfaction about a new product across different regions where it operates. The regions vary significantly in the number of customers. The stratified random sampling can be used to divide the population into separate regions.
| Regions | North | South | East | West | Total |
| Number of customers using the product | 10,000 | 20,000 | 15,000 | 5,000 | 50,000 (Population size) |
| Sample size of the stratum | 3,000 | 6,000 | 4,500 | 1,500 | 15,000 (Sample size) |
As discussed in the previous example, the company will calculate the sample size of the stratum using proportional stratified random sampling to collect data and perform the analysis.
Advantages and disadvantages of stratified random sampling
Stratified sampling has some key advantages and disadvantages, which should be taken into account before choosing it as a sampling technique for your research. Below is a brief list of advantages and disadvantages of stratified sampling.
Advantages of stratified random sampling
- Allows Comparisons: One of the major advantages of stratified sampling is that it allows precise comparisons between different subgroups. This allows a better understanding of the specific characteristics of the subgroups.
- Increased Precision: As stratified sampling ensures exclusive distribution of elements in each stratum, this reduces variability within each stratum hence decreasing the error in estimation and increasing the precision of the estimation. This is particularly beneficial when there are significant differences between strata.
- Ensures Representation: With stratified sampling, adequate representation of all relevant subgroups in the sample is possible. This is crucial for making accurate inferences about the entire population.
- Efficiency: By focusing on specific strata, stratified sampling can be more efficient and economical. It reduces the need for a very large sample size to achieve the desired level of precision.
- Reduced Bias: With proper implementation, stratified sampling reduces selection bias by ensuring that each stratum is appropriately represented.
Disadvantages of stratified random sampling
- Complexity: The major challenge of stratified sampling is that it is more complex and time-consuming compared to simple random sampling hence cannot be applied to all kinds of research. It requires comprehensive knowledge of the population to be able to identify and separate the strata.
- Strata Definition: Stratified sampling can only work efficiently when the strata are defined clearly, meaning every member of the population should strictly be a part of only a single subgroup. Poorly defined strata can lead to misrepresentation and inaccurate results.
- Population Knowledge: It also requires detailed knowledge of the population structure, which may not always be available and feasible. This includes knowing the size and characteristics of each stratum.
- Resource Intensive: It can be resource-intensive in terms of time, effort, and cost, especially if the population is large and complex. Hence, identifying and categorizing each member of the population into each subgroup can be challenging.
- Potential for Overlapping Strata: If individuals belong to more than one stratum, it can complicate the sampling process and lead to issues of overrepresentation leading to inaccuries and misinterpretation.
- Data Collection Challenges: Collecting data from each stratum can be logistically difficult, especially if strata are geographically dispersed or if some groups are less accessible than others.
Key takeaways
Stratified sampling involves dividing a population into distinct subgroups (strata) and sampling from each to ensure adequate representation. Stratified random sampling adds random selection within each stratum. This method is particularly useful for ensuring small or rare subgroups are represented, improving comparative analysis, and achieving specific research goals. It can be proportionate, reflecting the population’s structure, or disproportionate, to over-sample smaller groups. This approach enhances the accuracy and reliability of results compared to simple random sampling.
Frequently asked questions
1.Why is stratified sampling used?
Stratified sampling is used when subgroups need to be analyzed, where all groups should be adequately represented. This sampling method allows for increased precision and reliability of results.
2. How is stratified sampling different from simple random sampling?
In stratified sampling, the population is first divided into different subgroups based on the characteristics that are being studied. The strata are then used to draw samples using a random sampling method. While a simple random sampling method directly draws samples in a random fashion from the population without considering subgroups.
3. How do you determine the size of each stratum in stratified sampling?
The size of each stratum in stratified sampling can be decided disproportionately from the stratum or proportionally according to the size of the stratum relative to the total population.
4. When is stratified sampling most useful?
Stratified sampling is particularly useful when the population is large and has distinct subgroups. It aids in precision, reliability, and subgroup-specific insights.
5. What are the limitations of stratified random sampling?
The stratified sample has some key limitations such as:
- Complex design and implementation
- Challenging data collection
- Need for detailed population knowledge
- Requires more time and resources
- Risk of inaccuracies if strata are not defined properly
We hope the detailed information and examples provided in this article will help you get a better understanding of stratified sampling, including its types, uses and when to choose this sampling method to ensure accuracy and representativeness in your research.
R Discovery is a literature search and research reading platform that accelerates your research discovery journey by keeping you updated on the latest, most relevant scholarly content. With 250M+ research articles sourced from trusted aggregators like CrossRef, Unpaywall, PubMed, PubMed Central, Open Alex and top publishing houses like Springer Nature, JAMA, IOP, Taylor & Francis, NEJM, BMJ, Karger, SAGE, Emerald Publishing and more, R Discovery puts a world of research at your fingertips.
Try R Discovery Prime FREE for 1 week or upgrade at just US$72 a year to access premium features that let you listen to research on the go, read in your language, collaborate with peers, auto sync with reference managers, and much more. Choose a simpler, smarter way to find and read research – Download the app and start your free 7-day trial today!
