How do sociologists investigate society? » Sampling techniques - stratified and snowball

What you'll learn this session

Study time: 30 minutes

What sampling is and why it's important in sociological research
How stratified sampling works and when to use it
How snowball sampling works and when to use it
Strengths and limitations of both sampling techniques
Real-world examples of these sampling methods in action

Introduction to Sampling in Sociology

When sociologists study society, they rarely have the time or resources to talk to everyone in a population. Instead, they select a smaller group (a sample) that represents the larger population. The way researchers choose this sample is crucial to the quality and reliability of their findings.

Key Definitions:

Sampling: The process of selecting a subset of individuals from a larger population to study.
Population: The entire group that researchers want to learn about (e.g., all teenagers in the UK).
Sample: A smaller group selected from the population that researchers actually study.
Representativeness: How well the sample reflects the characteristics of the whole population.

Why Sampling Matters – A Real Example

In the 1936 US presidential election, a magazine called Literary Digest predicted that Republican Alf Landon would easily defeat Democrat Franklin Roosevelt. They based this on a huge sample of 2.4 million people. However, Roosevelt won by a landslide! The problem? Their sample was biased – they surveyed people from telephone directories and car registrations at a time when only wealthier Americans (who tended to vote Republican) had phones and cars. This famous polling disaster shows why how you sample is just as important as how many people you sample.

Stratified Sampling

Stratified sampling is a method where researchers divide the population into separate groups (called strata) and then select participants from each group. This ensures that key characteristics of the population are properly represented in the sample.

📈 How Stratified Sampling Works

In stratified sampling, researchers follow these steps:

Identify the important characteristics (e.g., age, gender, social class) that might affect your research
Divide the population into groups (strata) based on these characteristics
Calculate what percentage of the population each stratum represents
Select participants from each stratum in the same proportions as they exist in the population

💡 Example of Stratified Sampling

Imagine you want to study attitudes toward social media among students at a school with 1,000 pupils. The school has:

Year 7: 200 students (20%)
Year 8: 200 students (20%)
Year 9: 200 students (20%)
Year 10: 200 students (20%)
Year 11: 200 students (20%)

If you want a sample of 100 students, you would select 20 students from each year group to maintain the same proportions.

When to Use Stratified Sampling

Stratified sampling works best when:

You can clearly identify different groups within the population
These groups might have different views or experiences relevant to your research
You have access to information about the whole population to create accurate strata
You want to ensure smaller but important groups are properly represented

👍 Strengths of Stratified Sampling

Representativeness: Ensures all important subgroups are included in the correct proportions
Precision: Usually provides more accurate results than simple random sampling
Comparisons: Allows researchers to compare different subgroups
Smaller sample size: Can achieve reliable results with fewer participants

👎 Limitations of Stratified Sampling

Prior knowledge needed: Requires detailed information about the population
Time-consuming: Takes more time to organise than simple random sampling
Complexity: Can be difficult to decide which characteristics to stratify by
Access issues: May be hard to reach people in certain strata

Case Study: Health Survey for England

The Health Survey for England uses stratified sampling to gather information about the nation's health. First, they select a random sample of postcodes. Then, within each postcode, they select addresses in proportion to the population characteristics. This ensures they include people from different regions, age groups and socioeconomic backgrounds in the right proportions. The survey provides valuable data on health issues like obesity, smoking and mental health that helps shape government health policies.

Snowball Sampling

Snowball sampling is a technique where researchers recruit initial participants who then help identify and recruit additional participants. Like a snowball rolling downhill and getting bigger, the sample grows as more people join the study.

❄ How Snowball Sampling Works

In snowball sampling, researchers follow these steps:

Identify and contact a few initial participants who meet the criteria for the study
Ask these participants to recommend others who share similar characteristics or experiences
Contact these new people and invite them to participate
Ask these new participants to suggest more potential participants
Continue this process until you reach your required sample size or no new participants are suggested

💡 Example of Snowball Sampling

A researcher wants to study the experiences of young people who have dropped out of school. They might:

Contact a youth worker who introduces them to two former students
Interview these two young people about their experiences
Ask them if they know other school dropouts who might participate
Get introduced to three more participants
These three each suggest more people and the sample grows

When to Use Snowball Sampling

Snowball sampling is particularly useful when:

Studying hidden or hard-to-reach populations (e.g., homeless people, undocumented immigrants)
Researching sensitive topics where people might be reluctant to come forward
You have limited knowledge about or access to the target population
Trust is important for getting honest responses

👍 Strengths of Snowball Sampling

Access: Helps reach hidden or marginalised groups
Trust: Participants may be more open if referred by someone they know
Efficiency: Can quickly build a sample when no sampling frame exists
Insight: Can reveal social networks and connections

👎 Limitations of Snowball Sampling

Bias: Not representative of the wider population
Limited diversity: Sample may be restricted to certain social networks
No randomness: Participants are not randomly selected
Validity concerns: Findings cannot be generalised to the whole population

Case Study: Studying Drug Users

In 2018, researchers at the University of Manchester wanted to study the experiences of people who used synthetic cannabinoids (known as "Spice"). Because drug users are a hidden population who may fear legal consequences, the researchers couldn't simply advertise for participants. Instead, they used snowball sampling. They first contacted outreach workers at homeless shelters who introduced them to some initial participants. These participants then referred other users from their social networks. Through this chain of referrals, the researchers were able to interview 53 Spice users and gather valuable information about patterns of use, health impacts and reasons for using. This research helped inform support services and public health responses.

Comparing Stratified and Snowball Sampling

📝 Research Purpose

Stratified: Best for producing representative findings that can be generalised to the whole population.

Snowball: Best for exploratory research into hidden groups or sensitive topics where representativeness is less important than gaining any access at all.

👥 Population Access

Stratified: Requires good knowledge of and access to the population.

Snowball: Useful when population is hidden or hard to access through official channels.

📊 Data Quality

Stratified: Produces more reliable quantitative data suitable for statistical analysis.

Snowball: Often used for qualitative research where depth of insight matters more than statistical reliability.

Making the Right Choice

Choosing between stratified and snowball sampling depends on several factors:

What is your research question trying to find out?
Who is your target population and how easy are they to access?
Do you need statistically representative data or rich, detailed insights?
What resources (time, money, contacts) do you have available?

Remember that good sociological research matches the sampling method to the research aims. Sometimes, researchers even combine different sampling techniques in the same study to get the benefits of each approach.

Exam Tip!

In your exam, you might be asked to evaluate these sampling methods. Make sure you can:

Explain clearly how each method works
Give specific examples of when each would be appropriate
Discuss both strengths and limitations
Link your answer to practical research scenarios

Examiners love to see you apply your knowledge to real-world situations rather than just memorising definitions!

🧠 Test Your Knowledge!