💭 Subjective
Based on opinion and feelings, so it is biased. A researcher rates how "polite" a pupil seems just from their own impression. Another researcher might give a different rating.
Validity and Reliability » Validity, Subjectivity and Objectivity
What you'll learn this session
Study time: 30 minutes
Cambridge spec: 7.6.1, 7.6.2
- What validity means in research
- The difference between subjective and objective
- How to judge the validity of a new study
- Five ways to improve validity
Validity questions can come up in Section D of both papers. The idea is simple: is the study really measuring what the researcher wanted to measure?
🔒 Unlock Full Course Content
Sign up to access the complete lesson and track your progress!
Unlock This CourseWhat is validity?
Imagine a teacher wants to know how good pupils are at maths, but the test is full of very long reading passages. Pupils with weaker reading might score low even if they are good at maths. The test is measuring reading, not maths. That test would have poor validity.
Key terms:
- Validity: the extent to which the researcher is investigating what they aimed to investigate.
- Subjectivity: being impacted by personal opinions and feelings; biased.
- Objectivity: not being impacted by personal opinions and feelings; unbiased.
Validity is not about whether the results are interesting or whether the sample is big. It is about whether the study really tests its aim. If the aim is to find out whether a hot drink helps people stay alert, the study must measure alertness and nothing else.
Learn this
High validity: the study investigates what it aimed to investigate. Low validity: something else is being measured, or the results are twisted by opinion or by other factors.
Subjective or objective?
📏 Objective
Not based on opinion, so it is unbiased. A researcher counts how many times a pupil leaves their seat. Anyone counting would get the same number.
Subjective data lowers validity because the result might show the researcher's opinion, not the behaviour being studied. A stopwatch, a count, a score on a fixed test or a heart rate monitor are all objective measures.
This is not the same as the researcher being a bad person. Even a fair researcher can be affected by what they hope to find. That is why good research tries to remove opinion from how data is collected.
Judging validity in a new scenario
In the exam you will be given a study you have not seen before. Ask yourself three questions:
- Is the variable clearly defined so it can be measured?
- Is the measure objective, or does it depend on the researcher's opinion?
- Could anything else explain the results, or would participants answer or behave differently because of how they were asked or watched?
Worked example
A researcher wants to find out whether a new reading club makes pupils more confident at speaking. The teacher watches the pupils and writes down whether each one "seemed more confident" after the club.
Judgement: low validity.
Reason: "seemed more confident" is the teacher's opinion, so it is subjective and may be biased. The teacher may expect the club to work and so see what they hope to see.
Improvement: count how many times each pupil volunteers to speak in a ten-minute lesson. A count is an objective measure, so the data is no longer affected by the teacher's opinion.
How to improve validity
Here are five ways to improve validity. You have already met most of the ideas.
📝 Operationalise variables
Give each variable a clear, measurable definition. "Tiredness" becomes "number of yawns in ten minutes". Then the study measures what it aimed to measure.
🔒 Use controls
Keep other factors the same, such as the room, the time of day and the instructions. Then only the IV is likely to affect the DV, so the researcher is investigating what they intended.
⏱ Use objective measures
Collect data that does not depend on opinion, such as time taken, a count or a score. This removes bias from the researcher.
❓ Avoid leading questions
A question like "You enjoyed the lesson, didn't you?" pushes people to agree. Neutral wording lets participants give their true answer.
👁 Use covert observation
If people know they are being watched they may act differently. When the observer's role is hidden, people behave more naturally, so the behaviour recorded is the behaviour the researcher wanted to study. You met covert observation earlier. It raises ethical concerns, which are covered in the ethics lessons.
✅ Match the fix to the problem
Pick the improvement that fits the weakness in the scenario. Opinion-based rating? Use an objective measure. Vague variable? Operationalise it. Pushy question? Reword it.
Common mistakes
- Mixing up validity with sample size. A bigger sample does not make a study measure the right thing.
- Saying a study is "not valid" with no reason. Always say what is being measured wrongly, or where opinion or other factors get in.
- Writing "it is objective" just because numbers are used. If the researcher gives the number from their own opinion, such as a score out of ten for how friendly someone seems, it is still subjective.
- Suggesting an improvement that does not fit the problem.
Reliability, which is about consistency, is a different idea and is covered in the next lesson.
Exam-style question
A psychologist wants to find out whether doing a puzzle in a noisy room makes people more stressed. Each participant does a puzzle while loud music plays. Afterwards, the psychologist watches them and gives each person a score from 1 to 10 for how stressed they seem.
(a) Explain why this study may have low validity. [2 marks]
(b) Suggest one way the psychologist could improve the validity of this study. [2 marks]
Model answer
(a) The stress score is subjective (1) because it is based on the psychologist's own opinion of how stressed people seem, so it may be biased and may not measure real stress (1).
(b) Use an objective measure of stress, such as heart rate measured with a monitor (1), because this does not depend on the psychologist's opinion, so the study measures stress more accurately (1).
Exam tip
For an "improve" question, name a change that fits the exact problem in the scenario, then say why it helps the study measure what it aimed to measure.