ST8: Data Cleaning & Reliability
Understand how to clean data by dealing with missing values and outliers, and assess data quality through reliability, validity, bias and control groups.
Data Cleaning & Reliability
Understand how to clean data by dealing with missing values and outliers, and assess data quality through reliability, validity, bias and control groups.
Key Fact: Data cleaning means checking and correcting a data set before analysis — removing errors, dealing with missing values and identifying anomalies.
Key Fact: Missing data can occur when respondents skip questions or when measurements fail.
Key Fact: Options for missing data: exclude the record, use the mean/median for that variable, or investigate why it is missing.
Key Fact: An outlier is a value that is much higher or lower than the rest of the data.
Key Fact: Outliers should be investigated — they may be genuine extreme values or data entry errors.
Key Fact: If an outlier is a clear error (e.g. height recorded as 25 m), it should be corrected or removed.
Key Fact: If an outlier is genuine, it should usually be kept but its effect on results should be noted.
Key Fact: Reliability means the data or method would give consistent results if repeated.
Key Fact: Validity means the data or method actually measures what it is intended to measure.
Key Fact: Bias is a systematic error that skews results in a particular direction — it undermines validity.
Key Fact: A control group is used in experiments to provide a baseline for comparison — they do not receive the treatment.
Key Fact: Random allocation to control and treatment groups helps reduce bias in experiments.
📋 Key Vocabulary and Concepts
For Data Cleaning & Reliability, you must know:
- Data cleaning: The process of detecting and correcting errors, inconsistencies and missing values in a data set.
- Outlier: A data value that is significantly different from the other values in the set.
- Reliability: The extent to which a measurement or method gives consistent results when repeated.
- Validity: The extent to which a measurement or method actually measures what it is intended to measure.
- Bias: A systematic tendency for results to deviate from the true value in a particular direction.
- Control group: A group in an experiment that does not receive the treatment, used as a baseline for comparison.
❓ Practice Questions
Q: What is data cleaning?
Q: How should you deal with an outlier?
Q: What is the difference between reliability and validity?
Q: Why is a control group important in an experiment?
Q: Give an example of bias in data collection.
✅ Answers
- The process of checking a data set for errors, missing values and anomalies and correcting or removing them before analysis.
- Investigate it: if it is a data entry error, correct or remove it; if it is genuine, keep it but note its effect on results.
- Reliability means consistency (getting the same result on repetition); validity means the method measures what it is supposed to measure.
- It provides a baseline for comparison so the effect of the treatment can be measured, reducing the risk of attributing changes to the wrong cause.
- Asking about healthy eating only outside a gym — the sample is likely to include more health-conscious people than the general population.
🎯 Exam Tips
- Always investigate an outlier before removing it — never just delete data without a reason.
- When asked about data quality, comment on both reliability and validity.
- Bias can occur at any stage: sampling, question wording, data collection or analysis.
- Control groups are specific to experiments — do not mention them for surveys or observational studies.
- Explain the difference between a genuine outlier and an error using examples.
📝 Exam Technique
GCSE Statistics Exam Tips — Data Cleaning & Reliability:
1. For Data Cleaning & Reliability questions, show every step of your working clearly — method marks count even if the final answer is wrong
2. Check your answer makes sense in context (estimation, units, reasonableness)
3. Use correct mathematical notation and state formulae before substituting values
4. If a Data Cleaning & Reliability question asks you to 'prove' or 'show', write a logical chain of reasoning with a conclusion line
5. For problem-solving, identify the topic first, then recall the relevant method
1. For Data Cleaning & Reliability questions, show every step of your working clearly — method marks count even if the final answer is wrong
2. Check your answer makes sense in context (estimation, units, reasonableness)
3. Use correct mathematical notation and state formulae before substituting values
4. If a Data Cleaning & Reliability question asks you to 'prove' or 'show', write a logical chain of reasoning with a conclusion line
5. For problem-solving, identify the topic first, then recall the relevant method
⚠️ Common Errors
✗ Automatically removing all outliers from a data set ✓ Outliers should be investigated first — only remove them if they are clearly errors; genuine extreme values should usually be kept.
✗ Confusing reliability and validity ✓ Reliability is about consistency (repeat results); validity is about whether you are measuring the right thing. A method can be reliable without being valid.
✗ Thinking missing data should always be filled in with the mean ✓ Filling in with the mean can distort results — consider whether the missing data is random or systematic and whether excluding the record is better.
✗ Assuming a large sample guarantees unbiased results ✓ A large sample can still be biased if the sampling method is flawed (e.g. an online survey excludes people without internet access).
✍️ Model Answer
Full-Mark Response
A data set of students' heights in cm includes the values: 152, 148, 165, 15, 170, 155, 163, 159, 172, 145. Explain how you would clean this data set.
Step 1 — Identify anomalies: The value 15 cm is clearly an outlier. A height of 15 cm for a student is not realistic, so this is almost certainly a data entry error (likely meant to be 150 or 155). Step 2 — Investigate the outlier: Check the original data collection sheet. If the correct value can be found, correct it. If not, the value should be removed because it is clearly an error and would distort summary statistics (e.g. making the mean far too low). Step 3 — Check for missing data: Confirm all 10 records have been entered. If any records are incomplete, decide whether to exclude them or use an appropriate method to handle the missing values. Step 4 — Verify remaining values: The other heights (145–172 cm) are plausible for students, so they should be kept. After cleaning, the data set should be analysed and the removal of the error value should be documented.
📊 AO Deep Dive
Assessment Objective Analysis
AO1 (Knowledge & Understanding): Demonstrate knowledge and understanding of data cleaning & reliability, including data collection, presentation and calculation techniques relevant to Edexcel 1ST0 & AQA 8382.
AO2 (Application): Apply knowledge and understanding of data cleaning & reliability to interpret data, reason statistically and draw conclusions in context.
AO3 (Evaluation): Evaluate statistical methods and conclusions, assessing appropriateness, reliability, validity and bias through the statistical enquiry cycle.
📝 Exam Questions by Topic
- AQA Past Papers - Topic Questions -- AQA
- Edexcel Past Papers - Topic Questions -- Edexcel
- CCEA Past Papers - Topic Questions -- CCEA
🎬 Video Resources
Ready to ace your GCSE Statistics exams?
Get the best revision books and guides to boost your grades.