Sampling Simplified Study Notes

What is Sampling?

Sampling is a statistical process where you select a few items from a larger group called a population to gather information and make generalized findings about the entire population. Instead of studying every single member of a population, which is often impractical, expensive, or impossible, you analyze the sample and use the results to generalize findings to the larger group.

Elements of Sampling

These are the key components involved in the sampling process:

1. Population: The entire group of individuals, objects, events, or items that you are interested in studying. The population is defined by your research question.

Example: If you’re studying the opinions of university students in Jodan College, your population is all university students in the college.

• Sampling Frame: A list or other specification of the elements in the population from which the sample will be drawn. Ideally, the sampling frame should closely match the population.

Example: A list of all registered students at Moi University (if you’re studying students at that university). A database of customer email addresses. A telephone directory. The sampling frame is not always easy to obtain and is a constraint on the study

• Sample: The subset of the population that is actually selected for study.
  • Example: 500 students randomly selected from the Nairobi university’s student list.
• Sampling Unit: An individual member of the sampling frame.
  • Example: Each student listed in the Kenyatta university’s student directory.
• Variable of Interest: The characteristic or attribute that you are measuring or observing in your sample.
  • Example: Student’s GPA, their opinion on a new Presbyterian university policy, their major.
• Sampling Error: The difference between the characteristics of the sample and the characteristics of the population. This is unavoidable but can be minimized with appropriate sampling techniques and larger sample sizes.

Sampling Methods

There are two main categories of sampling methods: probability sampling and non-probability sampling.

Probability Sampling

These are sampling methods where every member of the population has a known and equal chance of being included in the sample. This allows for statistical inferences about the population. Probability sampling methods include:

1. Simple Random Sampling: Every member of the population has an equal chance of being selected. Selection is done randomly (e.g., using a random number generator).

For instant: Drawing names out of a pot. Using a random number table to select participants from a numbered list.

Merits of simple random sampling

1. This method is Unbiased. i.e. Every member has an equal chance of selection, minimizing bias.
2. It is Simple to understand or Conceptually easy to grasp.
3. If done correctly, it provides a good representation of the population.
4. It Allows the use of statistical methods to generalize findings to the population.

Demerits of simple random sampling

1. Requires a complete and accurate sampling frame i.e. it Can be difficult or impossible to obtain for large or dispersed populations.
2. Can be time-consuming and expensive especially for geographically dispersed populations, it might require travelling to many different locations.
3. May not be representative in small samples: Even though it’s unbiased, random chance can sometimes lead to samples that don’t accurately reflect the population especially if the sample size is small.
4. Systematic Sampling: Selecting members of the population at regular intervals e.g.  every 10th person on a list). The starting point is chosen randomly.

Example: Selecting every 5th customer who enters Naivas supermarket.

Merits systematic sampling

1. It is easier to conduct than simple random sampling, especially with large populations.
2. This method is Less time-consuming and less expensive than simple random sampling.
3. It can provide good representation if the population is randomly ordered.

Demerits systematic sampling

1. Requires a complete sampling frame
2. If the arrangement of the population list has a recurring pattern that coincides with the sampling interval, it can lead to systematic bias. For example, if every 10th house on a street corner is systematically different.
3. Less random than simple random sampling: Once the starting point is chosen, the rest of the sample is determined.

• Stratified Sampling: This method involves dividing the population into subgroups called strata(plural) based on shared characteristics e.g., age, gender, income e.t.c and then taking a random sample from each stratum(singular). This ensures representation from all subgroups. Strata are mutually exclusive to avoid repetition

Example: Dividing a population in Kibera slums into age groups 18-25, 26-35, 36-45, etc. and then randomly sampling from each age group proportionally to their representation in the population.

Merits Stratified Sampling

1. This methodguarantees that each stratum is represented in the sample, which is crucial when you want to analyze subgroups separately.
2. It leads to more precise estimates of population parameters compared to simple random sampling, especially when there is substantial variation between strata.
3. Allows for different sampling rates that is, a researcher can use different sampling rates in different strata to optimize the sample based on stratum size and variability.

Demerits Stratified Sampling

1. Requires knowledge of the population structure: You need to know how to divide the population into strata and have information about the size and variability of each stratum.
2. Can be more complex and time-consuming: Requires more planning and effort than simple random sampling.
3. Potential for misclassification: If individuals are incorrectly assigned to strata, it can introduce bias.
4. Cluster Sampling:

This method is characterized by dividing the population into clusters e.g., geographic areas, schools and then randomly selecting entire clusters to include in the sample. This is useful when the population is geographically dispersed or when it’s difficult to obtain a list of all individuals. Clusters ideally are mini reproductions of the overall population.

Example: Randomly selecting junior secondary schools in Thika city and then surveying all students within those selected schools.

Merits Cluster Sampling

1. It reduces travel costs and time, especially for geographically dispersed populations.
2. This method does not require a complete sampling frame of individuals: Only need a list of clusters.
3. Cluster Sampling is useful when it’s impossible or impractical to obtain a complete list of individuals in the population.

Demerits Cluster Sampling

1. Higher sampling error because individuals within clusters tend to be more like each other than individuals in the population as a whole.
2. If the clusters are not representative of the population, it can lead to biased results.
3. Requires careful cluster selection this means that it important to choose clusters that are as heterogeneous as possible.

• Multi-stage Sampling: Combines two or more of the above probability sampling methods. For example, you might use stratified sampling to divide a region into counties (strata) and then use cluster sampling to select schools within those counties.

This means that sampling is done into stages, the first sample is selected, then a second sample is selected from the first sample and a third sample is selected from second sample and so on, until a feasible sample is obtained.

Merits Multi-stage Sampling

1. Combines the benefits of different sampling methods  therefore it can tailor the sampling design to the specific needs of the research.
2. Efficient for large and complex populations i.e. it allows for cost-effective and practical sampling in situations where other methods are not feasible.
3. This method can be adapted to a wide range of research questions and population characteristics.

Demerits Multi-stage Sampling

1. More complex to plan and execute: Requires careful coordination of multiple sampling stages.
2. Higher sampling error: The more stages involved, the higher the potential for sampling error.
3. Requires expertise in sampling techniques: Needs a good understanding of the strengths and weaknesses of different sampling methods.

2. Non-Probability Sampling

These are methods where the selection of members is not based on random chance. These methods are often used for exploratory research or when probability sampling is not feasible. Results cannot be reliably generalized to the entire population. They include:

1. Convenience Sampling: Selecting participants who are easily accessible.

Example: Surveying students in your business statistics class. Asking people walking by on the street to participate.

Merits of convenience sampling

1. Easy and inexpensive: The simplest and least expensive sampling method.
2. Quick: Data can be collected quickly.
3. Useful for exploratory research: Can be helpful for generating initial insights or testing hypotheses.

Demerits of convenience sampling

1. The sample is unlikely to be representative of the population as a result of biasness
2. TheResults obtained through this method cannot be reliably generalized to the population.
3. Convenience samplingCannot be used to make statistical inferences about the population.
4. Purposive Sampling (Judgmental Sampling):

This method involves Selecting participants based on the researcher’s judgment of their knowledge or experience relevant to the study.

Example: Interviewing experts in finance field.

Merits of Purposive Sampling

1. It isEffective when you need to gather information from individuals with expertise or experience.
2. This MethodCan provide rich and detailed information.
3. Generally, it is less expensive than probability sampling methods.

Demerits Purposive Sampling

1. Subjective and prone to bias: The selection of participants is based on the researcher’s judgment, which can be biased.
2. Limited generalizability: Results cannot be reliably generalized to the population.
3. Requires expertise in the subject area: The researcher needs to have a good understanding of the population and the characteristics of interest.

Quota Sampling: This method is similar to stratified sampling, but participants are selected non-randomly to fill quotas for each stratum.

Example: Ensuring that a sample includes a certain number of individuals.

Merits of Quota Sampling

1. Ensures representation of subgroups i.e. It is Similar to stratified sampling, but without random selection.
2. Relatively inexpensive and quick: Easier to implement than stratified sampling.
3. Useful when probability sampling is not feasible: Can be used when a complete sampling frame is not available.

Demerits of Quota Sampling

1. Participants are selected non-randomly within each quota, which can introduce bias.
2. Results cannot be reliably generalized to the population.
3. Requires knowledge of population proportions i.e You need to know the proportions of different subgroups in the population.

• Snowball Sampling

In this method, existing participants recruit new participants from their networks. Useful for reaching populations that are difficult to access.

Example: Studying members of a rare disease support group.

Merits of Snowball Sampling

1. Useful for reaching hidden or hard-to-reach populations: Effective when it’s difficult to identify or contact members of the population.
2. Cost-effective: Can be a relatively inexpensive way to gather data.
3. Builds trust and rapport: Participants may be more willing to participate if they are referred by someone they know.

Demerits of Snowball Sampling

1. Bias towards initial contacts: The sample is likely to be biased towards the characteristics of the initial participants.
2. Limited generalizability: Results cannot be reliably generalized to the population.
3. Potential for ethical concerns: It’s important to protect the privacy of participants and ensure that they are fully informed about the research.