Collecting and classifying data
Studying this course? Keep your progress in one place.
Key facts
- Categorical data are words; numerical data are discrete (counted) or continuous (measured).
- A sample is chosen from the population; a census asks everyone.
- Sampling methods: simple random, systematic, stratified (in proportion) and convenience.
- Bias pushes results in one direction.
- Good questions are clear, neutral and give a time period.
Types of data
Data are categorical if described in words, and numerical if they are numbers that are either discrete (counted) or continuous (measured).
Categorical data are described in words or categories, such as bag colour or favourite sport.
Numerical data are numbers, and are either discrete or continuous.
- Discrete data can only take separate values, usually found by counting: number of siblings, goals scored, shoe size.
- Continuous data can take any value in a range, usually found by measuring: height, mass, time.
Continuous data are always rounded when recorded, but the true value could be anywhere in a range.
- 1
Data
what you collect
- 2
Categorical
described in words, e.g. bag colour
- 3
Numerical: discrete
counted, e.g. goals scored
- 4
Numerical: continuous
measured, e.g. height
| Categorical | Discrete | Continuous | |
|---|---|---|---|
| How it is found | Described in words | Counting | Measuring |
| Examples | Bag colour, favourite sport | Siblings, goals, shoe size | Height, mass, time |
Categorical
- How it is found:
- Described in words
- Examples:
- Bag colour, favourite sport
Discrete
- How it is found:
- Counting
- Examples:
- Siblings, goals, shoe size
Continuous
- How it is found:
- Measuring
- Examples:
- Height, mass, time
Which of these is continuous data?
Populations and samples
The population is everyone you want to know about, and a representative sample is a quicker way to learn about it than a census.
The population is every person or item you want to know about.
A sample is a smaller group chosen from the population. A census asks the whole population, but this is often too slow or expensive, so we use a sample.
A good sample is representative, meaning it has the same mix of people as the population. A larger sample is usually more reliable.
Population
Representative sample
Biased sample
Census
- Asks the whole population
- Often too slow or expensive
Sample
- Asks a smaller group
- Quicker, but must be representative
A school has 900 students. Asking every one of them is called
Sampling methods
There are four main methods: simple random, systematic, stratified and convenience.
There are four main ways to choose a sample.
- Simple random: every member has an equal chance of being chosen (random numbers or a draw).
- Systematic: choose every th member from a list, starting at a random point.
- Stratified: split the population into groups (strata) and take a random sample from each in proportion to its size.
- Convenience: choose people who are easy to reach. This is quick but often biased.
For students and a sample of , a systematic sample chooses every th student.
Simple random
- Every member has an equal chance
Systematic
- Every th member from a random start
Stratified
- Random sample from each group, in proportion
Convenience
- Easy to reach, often biased
Worked example: stratified sample
A club has members, of whom are adults. A stratified sample of is needed. How many adults and how many children?
- 1
Adults: .
- 2
Children: , so .
- 3
Check: .
A sample of 40 is taken from 800 students by choosing every th name from a list. What is ?
Bias
Bias pushes results in one direction, so they do not truthfully represent the population.
Bias means the results are pushed in one direction. To reduce it, use a random or stratified sample, a larger sample and neutral questions, and follow up non-responders.
- 1
Unrepresentative sample
asking people at only one place
- 2
Self-selection
only volunteers reply, such as an online poll
- 3
Non-response
many chosen people do not reply
- 4
Leading questions
suggest a particular answer
- 5
Biased results
pushed in one direction
Unrepresentative sample
- Asking people at one place
Self-selection
- Volunteers, such as an online poll
Non-response
- Many chosen people do not reply
Leading questions
- Suggest a particular answer
A website poll asks readers to vote on a new school uniform. Which source of bias is most likely?
Questionnaires and collecting data
A good question is clear, neutral and gives a time period, and answer boxes cover every answer without overlapping.
A good question is clear, neutral and gives a time period ("each school day"). Do not ask two questions in one.
Answer boxes should cover every possible answer, not overlap, and be easy to tick: None; Less than 1 hour; 1 to less than 2 hours; 2 hours or more. Include "none" or "other" where it is needed.
Primary data are collected by you (survey, observation, experiment). Secondary data were collected by someone else (a database or website).
Use a tally chart to record counts as you go. A short pilot survey with a few people shows up unclear questions before the real survey.
A good question
- Is clear
- Is neutral
- Gives a time period
- Asks only one thing
Good answer boxes
- Cover every answer
- Do not overlap
- Include none or other
Which set of answer boxes for "hours of sport per week" is best?
Try an exam question
A website poll asks readers whether the school should change its uniform. Give one reason the results may be biased and suggest one way to improve the sample.
[3 marks]
- [1]Only people who choose to vote take part (self-selection).
- [1]So the sample may not be representative of all students.
- [1]Use a random sample, for example random numbers on a list of all students.
That's the notes covered.
Carry on to the next subtopic.
Try a question
Which of the four pieces of information is continuous data?
1 mark, marked against the mark scheme
Which of the four pieces of information is continuous data?
1 mark, marked against the mark scheme