Welcome to the Statistical Problem-Solving Process!
Have you ever wondered how scientists predict election results, how companies decide which video games to make, or how wildlife experts count animals in a vast forest without catching every single one? They use the Statistical Enquiry Cycle!
In this chapter of your CCEA GCSE Statistics course (Unit 1 & Unit 2), you will learn how to plan an investigation, choose the right type of data, pick a fair sample, and collect information without falling into common traps and biases. Don't worry if this seems tricky at first—we will break down every concept step-by-step.
---1. The Statistical Enquiry Cycle
Statistics is not just about crunching numbers; it is a full 5-stage investigation process. Think of it as a repeating loop (an iterative cycle):
Stage 1: Hypothesis and Planning
Decide on a precise, testable question or statement. Plan what information you need and how you will get it.
Stage 2: Data Collection
Collect your information using surveys, experiments, or existing records. Make sure your methods are fair and unbiased.
Stage 3: Processing and Representing Data
Clean up the data, put it into tables, draw statistical charts/diagrams, and calculate summary values (like averages).
Stage 4: Interpreting Results
Look for patterns, trends, and relationships. Relate your findings directly back to your original hypothesis to see if it is supported or rejected.
Stage 5: Evaluating
Look back critically at your investigation. What went well? What were the limitations? How could you improve the process if you repeated it?
Memory Trick: Remember H-D-P-I-E (or "Have Delicious Pie In Evening") for Hypothesis, Data collection, Processing, Interpreting, and Evaluating!
Key Takeaway: The statistical enquiry process always starts with a clear question and ends with an honest evaluation of your methods.
---2. Planning and Formulating a Hypothesis
What is a Hypothesis?
A hypothesis is a precise, testable statement that you can investigate using data. It must never be a vague opinion.
Bad Example: "School lunches are bad." (This is an opinion and hard to measure accurately.)
Good Example: "Year 11 students spend more hours on homework per week than Year 8 students." (This is precise, measurable, and testable.)
Planning Considerations
Before jumping into data collection, a statistician must plan carefully:
• Variables needed: What exact measurements or answers do you need?
• Constraints: How much time, money, and resources do you have?
• Ethics & Sensitivity: Are the questions personal, sensitive, or offensive? Will participants' data be kept confidential and safe?
• Scope: Can you ask everyone (a census), or do you need a smaller group (a sample)?
Key Takeaway: A strong statistical investigation begins with a clear, measurable statement and a realistic plan that respects time, budget, and ethics.
---3. Classifications of Data
Data comes in many different flavours. Understanding the type of data you are dealing with helps you choose the right charts and calculations.
A. By Source (Where did it come from?)
• Primary Data: Data collected directly by you (the researcher) for your specific purpose.
Examples: You handing out questionnaires, taking measurements in a science lab.
Advantages: You know exactly how it was collected; it is up to date and tailored to your needs.
Disadvantages: Can be very time-consuming and expensive.
• Secondary Data: Data that already exists, collected by someone else for another purpose.
Examples: Census data from the Office for National Statistics (ONS), historical weather archives, published government reports.
Advantages: Cheap and quick to obtain; gives access to large datasets.
Disadvantages: May be out of date, contain unknown errors or bias, or not match your exact needs.
B. By Nature (What does it look like?)
• Qualitative (Categorical) Data: Non-numerical data consisting of words, labels, or categories.
Examples: Eye colour (blue, brown, green), favourite subject, car brand.
• Quantitative Data: Numerical data that can be counted or measured. This splits into two crucial types:
1. Discrete Data: Values that can only take specific, distinct, separate numbers (usually whole number counts).
Examples: Number of brothers and sisters (\(0, 1, 2, 3\)), shoe sizes (\(5, 5.5, 6\)), goals scored in a match.
Tip: If you count it, it is usually discrete!
2. Continuous Data: Values that can take any real numerical value along a continuous scale, limited only by the accuracy of the measuring tool.
Examples: Height of a person (\(165.4\text{ cm}\)), mass of an apple (\(120.35\text{ g}\)), time taken to run \(100\text{ m}\) (\(12.48\text{ s}\)).
Tip: If you measure it, it is continuous!
C. By Dimensionality (How many variables?)
• Univariate Data: Involves just one variable (e.g. recording only the heights of pupils).
• Bivariate Data: Involves two linked variables to see if there is a relationship or correlation (e.g. comparing height and arm span for each student).
Common Mistake to Avoid: Just because continuous data has been rounded (e.g. "age recorded to the nearest whole year" or "height to the nearest centimetre"), it is still fundamentally continuous in nature!
Key Takeaway: Identify whether your data is primary or secondary, qualitative or quantitative, and discrete or continuous before choosing your graphs and calculations.
---4. Population, Census, and Sampling
Key Definitions
• Population: The entire group of individuals, items, or animals that you are studying (e.g. all \(800\) students in a school).
• Census: Collecting data from every single member of the population.
Advantage: Completely accurate and unbiased view of the entire group.
Disadvantage: Very expensive, slow, and often impossible for large populations.
• Sample: A smaller subset selected from the population to represent the whole group.
• Sampling Frame: A complete list or database containing all members of the target population from which the sample is selected (e.g. the school register, an electoral roll).
Key Takeaway: A census surveys everyone, while a sample surveys a representative subset chosen from a sampling frame.
---5. Sampling Techniques
To make sure your sample gives a fair and accurate picture of the population, you need to use the right sampling method.
A. Simple Random Sampling
Every member of the population has an equal and known chance of being chosen.
Standard Step-by-Step Method (Essential for Exams!):
1. Number every item/person in the sampling frame from \(1\) to \(N\).
2. Generate random numbers using a random number generator, calculator, or random number table.
3. Select the items matching those numbers.
4. Ignore and discard any repeated numbers until the required sample size is reached.
B. Systematic Sampling
Selecting items at regular intervals from an ordered list.
Step-by-Step Method:
1. Calculate the interval value: \(k = \frac{\text{Population Size } N}{\text{Sample Size } n}\).
2. Choose a random starting point between \(1\) and \(k\).
3. Select every \(k^{\text{th}}\) item thereafter until you have your full sample.
Example: To pick a sample of \(n = 20\) from a list of \(N = 100\) students, \(k = \frac{100}{20} = 5\). Pick a random start between \(1\) and \(5\) (say, student \(3\)), then select students \(3, 8, 13, 18, 23\), and so on.
C. Stratified Sampling
When a population has distinct subgroups (called strata) such as different year groups or age categories, a stratified sample ensures each subgroup is represented in the exact same proportion as in the whole population.
The Stratified Sampling Formula:
$$\text{Number from stratum } i = \left( \frac{\text{Size of stratum } i}{\text{Total Population Size } N} \right) \times \text{Total Sample Size } n$$
Once you calculate the number needed from each stratum, you use simple random sampling within that group to select the individuals.
Worked Example:
A school has \(600\) students: \(200\) in Year 8, \(250\) in Year 9, and \(150\) in Year 10. You want a stratified sample of \(60\) students.
• Year 8: \(\left( \frac{200}{600} \right) \times 60 = 20\) students
• Year 9: \(\left( \frac{250}{600} \right) \times 60 = 25\) students
• Year 10: \(\left( \frac{150}{600} \right) \times 60 = 15\) students
• Check total: \(20 + 25 + 15 = 60\) students. Always double-check your numbers add up to the total sample size!
D. Quota Sampling (Non-random)
The researcher is given a quota of people with specific characteristics to interview (e.g. \(20\) males over \(50\), \(20\) females under \(30\)). No sampling frame is used; the researcher simply stops people who fit the profile until the quota is filled.
E. Opportunity / Convenience Sampling (Non-random)
Choosing people who are easiest to reach (e.g. standing outside a shopping centre on a Tuesday morning and interviewing passers-by). While quick and cheap, this method is biased because it excludes people who are at work, school, or elsewhere.
Key Takeaway: Random, systematic, and stratified sampling are probability methods that minimise bias. Quota and opportunity sampling are non-random and prone to bias.
---6. Estimating Populations: Capture-Recapture Method
How do biologists estimate the number of fish in a lake? Catching them all is impossible! Instead, they use the Capture-Recapture method.
The Formula
$$N = \frac{M \times n}{m}$$
Where:
• \(N\) = Total estimated population size
• \(M\) = Number of individuals caught, marked, and released in the first sample
• \(n\) = Total number of individuals caught in the second sample
• \(m\) = Number of marked individuals found in the second sample
Worked Example
An ecologist catches \(50\) field mice (\(M = 50\)), marks their tails with non-toxic paint, and releases them. A week later, she catches \(40\) mice (\(n = 40\)) and finds that \(8\) of them are marked (\(m = 8\)).
$$N = \frac{50 \times 40}{8} = \frac{2000}{8} = 250\text{ mice}$$
Crucial Assumptions (Frequently Tested in CCEA Exams!):
For this estimate to be valid, we must assume:
1. The population is closed (no births, deaths, immigration, or emigration between samples).
2. The marks do not rub off, fade, or wash away.
3. The marks do not harm the animals or make them easier for predators (or researchers) to catch.
4. The marked animals have enough time to mix randomly back into the population.
Key Takeaway: Capture-recapture estimates wild populations using a simple ratio, but its accuracy depends heavily on biological assumptions.
---7. Designing Questionnaires and Avoiding Bias
Writing a great questionnaire takes skill. Poorly written questions lead to inaccurate or misleading data.
Rules for Designing Great Survey Questions
1. Make options Exhaustive and Mutually Exclusive:
• Mutually exclusive: Response boxes must not overlap.
Wrong: [ ] \(0 - 5\) [ ] \(5 - 10\) [ ] \(10 - 15\) (Where does a person with \(5\) tick?)
Right: [ ] \(0 - 4\) [ ] \(5 - 9\) [ ] \(10 - 14\)
• Exhaustive: Cover all possibilities (include options like "None", "Other", or "Prefer not to say").
2. Avoid Leading or Biased Questions:
Do not push the respondent toward a particular answer.
Wrong: "Don't you agree that school uniform is uncomfortable?"
Right: "To what extent do you agree or disagree that school uniform is comfortable?"
3. Use Specific Timeframes:
Avoid vague words like "often" or "recently". Instead, ask: "How many times have you visited the cinema in the last 7 days?"
What is a Pilot Survey?
A pilot survey is a small-scale trial run conducted on a small test group before launching the full investigation.
Purpose: To spot confusing questions, test how long it takes, and fix any errors before spending time and money on the main survey.
Types of Bias to Watch Out For
• Sampling / Selection Bias: When your sample does not accurately represent the target population (e.g. asking only gym members about their exercise habits to find the national average).
• Response Bias: When respondents give untruthful or socially desirable answers because they feel embarrassed or want to impress the interviewer.
• Non-response Bias: When people who refuse or fail to answer the survey have systematically different opinions or traits from those who do answer.
Key Takeaway: Good questionnaires use clear, non-overlapping tick boxes, avoid leading language, and are pre-tested using a pilot survey to eliminate bias.
---8. Quick Summary & Exam Checklist
Before you tackle exam questions on this topic, check that you can:
✔ List the 5 stages of the Statistical Enquiry Cycle in order (Hypothesis \(\rightarrow\) Data Collection \(\rightarrow\) Processing \(\rightarrow\) Interpreting \(\rightarrow\) Evaluating).
✔ Classify any data correctly as primary/secondary, qualitative/quantitative, and discrete/continuous.
✔ Describe the full 4-step process for generating a simple random sample.
✔ Calculate sample sizes for stratified sampling using the formula.
✔ Apply the capture-recapture formula (\(N = \frac{M \times n}{m}\)) and state its key assumptions.
✔ Identify flaws in questionnaire questions (overlapping boxes, vague timeframes, leading questions) and write improved versions.
✔ Explain the purpose of a pilot survey.