In this you will learn How Stratified Sampling Works? Steps to perform Stratified Random Sampling, Applications, Pros and Cons. Also you will get to download PDF.
When conducting research, sometimes it is impractical or impossible to study an entire population. In these cases, researchers use a method called sampling. Sampling is the process of selecting a group of people (the sample) from the larger group (the population) to participate in the study. There are different methods of sampling, but one common method is stratified sampling.
In social science research, stratified sampling is a method for selecting a sample population in which the members are grouped into strata, and a random sampling method is then used to select a proportional number of members from each stratum. This approach is often used when it is important to ensure that the different strata in the population are represented in the sample in order to avoid bias.
In stratified sampling, the population is divided into subgroups or strata, and a sample is taken from each stratum. The goal of stratified sampling is to improve the representativeness of the sample by ensuring that each stratum is represented in proportion to its size in the population.
Stratified sampling can be used when there is heterogeneity within the population, and it can be particularly helpful when there is a limited amount of data available.
Example of Stratified Sampling
Suppose we want to study the relationship between income and job satisfaction. We might use stratified sampling to divide the population into subgroups based on income level, and then select a random sample from each subgroup. This would ensure that we have a representative sample of both low-income and high-income individuals.
How Stratified Random Sampling Works?
It is possible for a researcher to discover that the population size is too large to do research on a set of things with similar characteristics.
An analyst may choose a more suitable course of action by choosing a small subset of the population in order to save time and money.
A sample size, or small group, is a portion of the population used to represent the total population. The stratified random sampling method is one of many methods for choosing a sample from a population.
In stratified random sampling, the entire population is divided into uniform groups known as strata (plural for stratum). Then, a random sample is chosen from each stratum.
You must be able to divide your population into mutually exclusive and exhaustive subgroups in order to employ stratified sampling. This implies that there is only one distinct subgroup into which every member of the population may be categorized.
When you think that subgroups will have distinct mean values for the variable(s) you’re examining, stratified sampling is the best option among the probability sampling techniques.
Consider a researcher in academia who is curious about how many BA students in 2022 obtained job offers within four months of graduating. The researcher will soon discover that the number of BA graduates for the year was close to 100,000. They might choose to conduct a survey with a straightforward random sample of 40,000 recent grads. Even better, they could stratify the population and select a representative sample at random from each stratum. To do this, they would divide the population into groups according to the gender, age range, race, nation of origin, and professional experience. A random sample is drawn from each stratum in a quantity proportional to its size relative to the population. The strata’s subsets are then combined to create a random sample.
Steps to Perform Stratified Sampling
As with other probability sampling techniques, you should start by stating the population from which your sample will be drawn.
Step 1 – Define population and subgroups
✔ Selecting traits for Stratified Sampling
Additionally, you must decide on the trait that will be used to split your groups. The classification of each subject to each subgroup should be transparent and clear because each person in the population can only be assigned to one subgroup.
✔ Stratifying based on multiple factors
If you can clearly identify each subject as belonging to only one subgroup, you can opt to stratify by several separate attributes at once. In this instance, multiplying the numbers of strata for each characteristic results in the total number of subgroups.
Example: stratifying based on multiple traits
If you were to stratify by race and gender identity, for instance, using five groups for the former and two for the latter, you would have a total of 10 groups (5 x 2).
All of the individuals in your demographic are recent college undergraduates who attended the university. You will stratify based on degree and gender identification.
Step 2 – Separate the population into strata
The next step is to compile a list of every person in the population and stratify them.
You must make sure that each stratum contains the complete population and that they are all mutually exclusive (i.e., there is no overlap between them).
Example: Stratifying the population
Every graduate’s name, gender identity, and qualification are listed on the list. You can categorize using this list based on two characteristics: gender identification (male, female), and qualification (bachelor’s, master’s, PhD, none). By combining these traits, you can create a total of eight groups. Each graduate must be put into one specific category.
|Gender||■ Male |
|1. Male bachelor graduates|
2. Female bachelor graduates
3. Male Master graduates
4. Female Master graduates
5. Male PhD graduates
6. Female PhD graduates
7. Male having no qualification
8. Female having no qualification
Step 3 – Choose sample size for each stratum
You must first choose if you want a proportionate or disproportionate sample.
In proportional sampling, the sample size for each stratum corresponds to the proportion of the subgroup in the total population.
Subgroups that are underrepresented in the overall population will likewise be underrepresented in the sample, such as rural people, which in most nations make up a smaller percentage of the population.
In disproportionate sampling, each strata’s sample sizes are larger than the proportion of that strata in the whole population.
This approach might be used if you want to research a highly underrepresented population but your sample size is too small to make statistical inferences.
Size of Sample
You can choose your overall sample size after that. This should be a sizable enough sample size to allow you to make statistical inferences about each category.
You can use a sample size calculator to calculate the appropriate figures if you know your desired margin of error, confidence level, projected size, and standard deviation of the population you’re dealing with.
Example: sample size
You choose to utilize disproportionate sampling because you need to make sure your sample of master’s grads is huge enough.
Although master students make up a relatively tiny fraction of the total student body, your sample contains roughly one-third graduates of bachelor’s, one-third graduates of master’s, and one-third graduates of doctorate programs.
Step 4 -Sample at random from each stratum
Finally, you should sample from within each stratum using a different probability sampling technique, such as simple random or systematic sampling.
If done correctly, the inherent randomization in such procedures will let you get a sample that’s typical of that specific segment.
Example: Random Sampling
You pick individuals using simple random sampling, taking an approximately equal sample size from each of your nine groups.
In order to answer your question, you can then gather information from each member of your sample regarding their incomes and employment history.
Benefits of Stratified Sampling
Most large-scale surveys utilize stratified sampling because of its many benefits, some of which are listed below:
▶ Subpopulation Estimation
Such subpopulations should be treated as strata when the estimates of the population characteristics are required for both the population as a whole and its various subpopulations.
For instance, the government might be interested in estimating national and provincial unemployment rates in a national survey on unemployment. Each province can be viewed as a strata in this situation.
▶ Representativeness of sample
The formation of strata and the distribution of samples among them can be done through stratified sampling in such a way that the sample can represent the population with regard to the variables being studied.
For instance, a simple random sampling without replacement (SRSWOR) sample from the entire school may not be representative if we want to choose a sample of pupils from a school that represents the various races in South Africa. As opposed to an SRSWOR sample from the entire school, in this case stratified sampling employing several racial groupings as strata is anticipated to yield a more representative sample.
▶ Administrative ease
The organization conducting the survey may stratify the population in order to efficiently supervise the survey, for example, by appointing different supervisors to conduct the survey for each of the strata separately.
▶ Better Data Quality
Better data quality can be attained by assigning various investigator types to various strata. For instance, investigators fluent in the native tongue may be sent to rural areas, whereas investigators fluent in English may be more useful in urban areas.
By creating strata that are homogeneous with regard to the attribute under study, stratification may improve the estimates’ efficiency. The efficiency of the estimators may be increased by using sampling plans that are suitable for the particular strata.
▶ Cost Effective
By dividing a large population into smaller groups with similar members rather than sampling every member of a larger population, researchers can reduce the costs associated with their research.
Limitations of Stratified Sampling
▶ Sampling Error
When the sample fails to precisely reflect the population as a whole, sampling errors might happen. The researcher would have to start the sampling procedure over if this happened.
▶ Requires Planning
Researchers must make sure that every person in the population belongs to just one stratum and that the total population is represented by all the strata. Simple random sampling does not require as much additional planning and data gathering as this.
▶ Differences among the population
If there are too many differences among the population or not enough data available, it is impossible to divide the population into subgroups.
When participants belong to several subgroups, overlapping can become a problem. Multiple subgroup members are more likely to be chosen when simple random sampling is used. The outcome can be a distorted or false reflection of the people.
Applications of Stratified Sampling
Stratified Sampling is useful when
■ a population’s complete population is not available to researchers.
■ examining the earnings of various populations or the earnings of various jobs across a country.
■ analyzing life expectancy, population demography, or election polls.
■ more efficient and affordable than other sample techniques.
■ The creation of strata aids in group organization when population samples differ greatly.