How Stratified Random Sampling Works?

Table of Contents

What Is Stratified Random Sampling?

Stratified random sampling is a method where the population is separated into smaller subgroups, termed strata, and from each subgroup, a sample is selected. Strata are designed on the basis of the common features or qualities of the members. These characteristics may include income, educational background; gender, etc. This sampling method is also called proportional random sampling or quota random sampling.

Key Points

A stratified random sample is done on the basis of the division of the population into homogenous groups known as strata.
This type of sampling helps the researchers obtain a sample from the population that best represents the population.
Proportional stratified random sampling helps to take random samples from the clusters as a fraction of the population. In disproportionate sampling, there are no proportional strata.
This sampling technique is different from simple random sampling.

Steps to conduct stratified random sample selection

The selection of stratified random samples involves the following steps:

The first step is defining the target audience.
Secondly, the recognition of variables and the identification of the number of strata to be used in research. It is worth remembering that the strata should be formed according to the objectives of the research.
Thirdly, to either use a prevailing sampling frame or generate a frame that comprises all the data about the variables to be stratified for all the essentials of the target listeners.
Fourth, make changes to the sample frame on the basis of absence of coverage, over-coverage, and/or grouping.
The fifth step is to make sure that every stratum is unique and covers each member related to population. Inside the stratum, there must be the least difference, and among the strata, there should be a significant difference.
Sixth, to assign a distinctive number to every element.
In the seventh step, the process of picking up the size of every stratum according to the researcher’s needs begins. The arithmetic scattering amongst total elements in total strata will determine the type of sampling to be executed. On this basis, the stratified random sampling will be either proportional or disproportional.
Lastly, there is a random selection of samples from each stratum for the formation of the sample. Each stratum should have a representation in the sample. On the other hand, if two components from every section are selected, quickly calculate the error margins of the calculation of the collected data.

Kinds of stratified random sample selection

The two types of stratified random sample selection techniques are discussed below.

1. Proportionate sampling

In this type of sampling, there is a direct proportion of sample size to the population size of the entire population of the strata. It indicates that the strata sample has the same sampling fraction.

Proportionate stratified random sample selection Formula: n_h = ( N_h / N ) * n

n_h= Sample size for the stratum

The N_h= Population size for the stratum

N = Size of the whole population

The n = Size of the whole sample

2. Disproportionate Sampling:

It is an approach where the sample size drawn from each stratum or level is not in proportion to the size of the stratum or level in the whole population. With too much sampling, there will be different sampling fractions for each stratum. This sampling method can be successful if the researcher precisely allocates the fraction. If the allotted fractions are not accurate, there may be bias in the results because of overrepresentation or underrepresentation of the strata.

Stratified random sample selection example

The stratified random sample selection is applied for analyzing the relationship between two or more strata. So it is important to compute the sections (strata) prior to computing the value of the sample.

Let us take the example of 100 (N_h) students from the population of 1000 (N) students who were asked to highlight their favorite subject. The sample will consist of students from different classes, such as 5^th, 6^th, 7^th, 8th, and 9^th. 8^th grade students will have different subject preferences as compared to 9th grade students. For getting precise results, each step is divided into various strata. The following table shows the students’ numbers in each class.

Class	Number of students (n)
5^th	150
6^th	250
7^th	300
8^th	200
9^th	100

For calculating the sample size for each class, the following formula is used:

Stratified Sample (n₅) = 100 / 1000 * 150 = 15

Stratified Sample (n₆) = 100 / 1000 * 250 = 25

Sample (Stratified) (n₇) = 100 / 1000 * 300 = 30

Stratified Sample (n₈) = 100 / 1000 * 200 = 20

Stratified Sample (n₉) = 100 / 1000 * 100 = 10

Purpose of stratified random sample selection

The researcher, while conducting his research, may find the population too large to reach and collect data from it. For this purpose, he or she selects a sample from that population, which is referred to as the sample size. This selection of the sample takes place in many ways, one of which is the stratified random sampling technique.

In stratified random sampling, the population is divided into many homogenous groups known as strata on the basis of their demographics, such as age, gender, race, etc. For example, a researcher wants to know how many students, after graduating from the university in a particular year, got jobs three months after completing their graduation.

The researcher will calculate the number of graduates in that particular year. After that, he or she will divide the population into subgroups on the basis of age, gender, race, nationality, or educational background. From each subgroup, the researcher will select a sample size proportional to the size of the stratum compared to the population. Then the samples from each stratum are pooled for the purpose of forming a random sample. Lastly, the results obtained will be compared with each other to find the differences among them.

When should I use stratified random sampling?

The stratified random sampling technique remains used in situations where researchers are keen to identify the diverse subgroups of the total population. For example, if the researcher is interested in knowing the difference in anxiety levels among male and female students, stratified random sampling is used.

Stratified random sample selection is also used in situations where the population has different subgroups and the goal is to find the differences in those subgroups.

How are the strata chosen?

The strata are chosen when the population is comprised of members with different characteristics. In this process, the size of each stratum is figured out according to the researchers’ requirements. The distribution amongst all the components in every stratum will determine the type of sampling to be implemented. On this basis, the stratified random sampling will be either proportional or disproportional. Furthermore, there is a random selection of samples from each stratum for the formation of the sample. Each stratum should have a representation in the sample; however, if two segments from every stratum (part) are selected, rapidly compute the error margins of the calculation of the collected data.

Advantages of stratified Random Sample selection

There are various advantages to this type of sampling. One major advantage of this sampling technique is its identification of significant population features in the sample.
The subgroups are formed with the help of stratified random sampling.
This technique gives greater precision in the sample selection as well as the findings.

Disadvantages of Stratified Random Sample selection

This sampling process is not used in every research study.
One has to meet several conditions before its use is effective and proper.
There is a need for identification and classification of every member of the population being studies in a subgroup, which is a difficult task. The resultant population list can be challenging.
In cases when there is overlapping of the subjects due to their falling in more than one group, there can become issue to deal with.