Sampling and Bias (Edexcel GCSE Maths): Revision Notes
Sampling and bias
When conducting statistical investigations, we need to gather information from groups of people, objects, or events. However, it's rarely possible to collect data from absolutely everyone we're interested in studying. This is where sampling comes in - but choosing the right sample is crucial for getting reliable results.
Sampling is fundamental to statistical research because studying entire populations is often impractical due to time, cost, and logistical constraints. The quality of your sample directly impacts the validity of your conclusions.
Understanding populations and samples
What is a population?
A population refers to the complete group that you want to investigate or learn about. This could be anything - all the students in your school, every penguin in Antarctica, or all the cars on British roads. The population represents everyone or everything that your research question applies to.
What is a sample?
Since studying entire populations is often impossible due to their size, time constraints, or cost, we select a smaller group called a sample. This sample should ideally give us insights that we can apply to the whole population.
The key challenge is ensuring that your sample accurately reflects the characteristics of the larger population you're studying.
Requirements for representative sampling
For your sample to provide reliable insights about the population, it must be representative. This means it should mirror the key characteristics of the entire population you're investigating.
Two essential requirements
Random sampling: Every member of the population must have an equal opportunity of being selected for your sample. This prevents you from accidentally favouring certain types of people or items, which could skew your results.
Adequate sample size: Your sample needs to be large enough to capture the diversity within your population. A sample that's too small might miss important variations or be heavily influenced by unusual cases.
Identifying bias in sampling methods
Bias occurs when your sample doesn't properly represent the population you're studying. This can happen in various ways and can seriously affect the reliability of your conclusions.
Key questions to spot bias
When evaluating any sampling method, consider these crucial factors:
Key Questions to Ask:
- When is the sample being taken?
- Where is the sampling taking place?
- How are participants being selected?
- How many people are included in the sample?
Common sources of bias
Exclusion bias: When certain groups are systematically left out of your sample. This might happen due to the location, timing, or method of data collection.
Size bias: When your sample is too small to represent the population adequately, making it vulnerable to random variations.
Selection bias: When the method of choosing participants favours certain types of people over others.
Demographic bias: When your sample doesn't reflect the age, gender, interests, or other characteristics of the broader population.
Understanding these types of bias helps you identify potential problems in your sampling approach before they affect your results.
Real-world examples of sampling problems
Train station survey example
Worked Example: Train Station Survey
Scenario: Someone wants to understand how often people travel by train, so they survey people at a train station one morning.
Problems identified:
- The sample excludes people who never use trains, so it will overestimate train usage in the general population
- Surveying only at one station and one time of day means the sample might not represent different types of train users or travel patterns
Type of bias: Exclusion bias and selection bias
School orchestra example
Worked Example: School Orchestra Survey
Scenario: A student wants to know if pupils at their school would like more music lessons. They decide to survey 10 members of the school orchestra out of 800 total pupils.
Problems identified:
- Sample is far too small to represent 800 students reliably
- Orchestra members are likely to be more enthusiastic about music than the average student, creating a bias towards supporting more music lessons
Type of bias: Size bias and demographic bias
Strategies for better sampling
Improving randomness
Try to give every member of your population an equal chance of selection. This might involve using random number generators, systematic selection methods, or ensuring your sampling location and timing don't favour particular groups.
Random sampling doesn't mean haphazard sampling. It requires careful planning to ensure true randomness and equal opportunity for selection.
Ensuring adequate size
Larger samples generally provide more reliable results, but the exact size needed depends on your population size and the diversity within it. Consider what would give you confidence in your conclusions.
Minimising exclusion
Think carefully about who might be left out by your sampling method. Are there groups that your approach systematically misses? How could you adjust your method to include them?
Consider the practical barriers that might prevent certain groups from participating in your sample, such as time of day, location accessibility, or communication methods.
Key Points to Remember:
- A population is the entire group you want to study, while a sample is the smaller group you actually collect data from
- Representative samples must be both random and appropriately sized for reliable results
- Bias occurs when your sample doesn't properly reflect your population - ask yourself when, where, and how you're sampling
- Common bias sources include excluding certain groups, inadequate sample size, and non-random selection methods
- Always consider who might be left out of your sample and how this could affect your conclusions