IB Psychology SLTopic 5 — Research DesignPaper 1 & 2Core idea~11 min read
Validity: Measuring What You Meant To
Validity asks a harder question than reliability. Not “would I get this again?” but “is this true, and true of anyone other than the people in the room?” Almost every evaluation point you will ever write in psychology is a validity point wearing a different hat.
📘 What you need to know
Validity = the findings are true to life and measure what they were meant to measure.
Internal validity: the change in the DV really came from the IV, not from something else.
Confounding variables are the enemy of internal validity.
External validity: the findings apply beyond the research setting.
Ecological validity is a type of external validity, about how realistic the task and setting are.
Temporal validity: do the findings still hold in a different era?
Construct validity: does the measure capture the idea it claims to? Predictive validity: does it forecast later behaviour?
The four questions validity asks
When a study reports a result, there are four separate ways it can be wrong, and psychology has a name for each. Learning the names is most of the work.
Construct and predictive validity sit under the same roof but do not belong to either branch neatly. Treat them as two more questions to ask of any measure.
Internal validity
Internal validity is about the inside of the study. If the noisy group remembered fewer words, was it the noise? Or was it that they were tested at 4pm while the quiet group was tested at 9am? Anything that changes alongside the IV and could explain the result is a confounding variable, and it destroys internal validity.
The defences are the ones you already know: standardised procedures, controlled conditions, random allocation, and using the same materials throughout. When a study is well controlled, the conclusions drawn from it can be trusted, because there is nothing else left to blame.
The strongest phrasing for an exam: “because the two groups differed only in the IV, the difference in the DV can reasonably be attributed to it. This gives the study high internal validity.”
External validity
External validity asks whether the finding travels. Three things can stop it.
The task and setting. If the task is artificial and the room is sterile, behaviour there may not resemble behaviour anywhere else. That is ecological validity.
The people. If the sample was 60 psychology undergraduates from one university, the results may not describe anyone else.
The time. Studies reflect the world they were run in. That is temporal validity.
This is the same trade-off you met with lab and field experiments, now with its proper names attached.
A useful sentence to keep: high control buys internal validity and spends external validity. If you can say which one a study bought and which one it sold, you are evaluating properly.
Ecological validity, more carefully
Students often assume ecological validity is only about location. It is really about whether the task resembles something people actually do, and whether their engagement with it is real.
A study can take place in a laboratory and still have decent ecological validity, if participants are genuinely engaged in what they are doing. And a study in a real street can have poor ecological validity if the task is bizarre. Location is a clue, not the answer.
Temporal validity
Some findings age badly, because the world they described has changed. Research on family life carried out when almost every household looked the same will not describe a world of single parents, blended families and same-sex parents. Research on obedience or conformity reflects the social climate it was run in.
Use this carefully. “It is old, so it is invalid” is a weak point. “The social conditions it depended on have changed, so we should not assume the same result today” is a strong one. Name the condition that changed.
Construct and predictive validity
Construct validity matters most for abstract things: intelligence, mood, empathy, depression. You cannot see them, so you build a measure and then have to argue that the measure really captures the idea. A questionnaire that mostly measures how tired someone feels is not a valid measure of depression, however consistent it is.
Predictive validity asks a simpler question: does the score forecast what happens later? A measure of infant attachment has predictive validity if it tells you something about relationships years afterwards. This is what makes psychological measures useful in schools, clinics and workplaces.
Type
The question it asks
Threatened by
Protected by
Internal
Did the IV cause the DV?
Confounding and extraneous variables
Control, standardisation, random allocation
Ecological
Is the task and setting realistic?
Artificial tasks and sterile settings
Real settings, meaningful tasks
Temporal
Does it still hold today?
Social change since the study
Replication in the present day
Construct
Does it measure the idea itself?
Vague or badly defined concepts
Clear definitions, unambiguous tasks
Predictive
Does it forecast later behaviour?
Scores that relate to nothing
Follow-up studies over time
Worked examples
WORKED EXAMPLE
Identify the threat to validity
In a study of memory, the quiet condition was run on Monday morning and the noisy condition on Friday afternoon. Identify the problem and explain its effect. [3]
Step 1: Find what changed alongside the IV
Time of week and time of day changed with the condition.
Step 2: Name it properlyA confounding variable — tiredness on Friday afternoon could explain lower recall by itself.
Step 3: State the consequence
The researcher cannot say the noise caused the difference.
Internal validity is reducedthe fix is simple: run both conditions at the same time of day, or alternate them
WORKED EXAMPLE
Evaluate the validity of a finding
A lab study of helping behaviour uses 48 first-year psychology students who watch a filmed emergency and press a button if they would help. Evaluate the validity of the conclusions. [6]
Internal validity: strong
Controlled conditions and a standardised film mean the IV is the only thing differing between conditions.
Ecological validity: weakPressing a button about a film is not helping anyone. The behaviour measured is not the behaviour of interest.
Population validity: weak
48 psychology students are not a cross-section of people, and may already know the theory.
Construct validity: questionable
Is a button press really a measure of helping, or of stated intention?
High internal, low external and construct validitystructure the answer by type — it makes the marking obvious
💡 Exam tip
Never write “low validity” alone. Always name the type and explain it in the context of the study.
Internal and external validity usually pull against each other. Saying so shows real understanding.
For ecological validity, comment on the task as well as the setting.
Demand characteristics reduce validity, not reliability. Be precise about which.
When suggesting improvements, match the fix to the type: control fixes internal, realism fixes ecological, replication fixes temporal.
Construct validity is the one most students forget. Using it well stands out.
⚠️ Common mix-up
Validity and reliability. Consistency is not accuracy.
Ecological validity means “outdoors”. It does not. It means realistic task and genuine engagement.
Extraneous vs confounding. Extraneous variables are anything else that could affect the DV; confounding ones actually vary with the IV.
Thinking a big sample gives internal validity. Sample size helps generalisability, not causation.
Calling all old studies temporally invalid. Only if the conditions they relied on have actually changed.
Treating validity as one thing. There are several, and questions usually target one.
Up next: Generalisability: Who Else Do the Results Apply To? — taking external validity apart and asking exactly how far a finding reaches.
Want this explained one-to-one?
Book a free session with an experienced IB Psychology tutor and get your trickiest topics made simple.
