IB Psychology HLTopic 5 — Research DesignPaper 3 & IACore idea~9 min read
Generalisability: Who Else Do the Results Apply To?
Every study is done on a handful of people, and every conclusion quietly claims to be about far more than them. Generalisability is the question of whether that jump is allowed — and it is one of the easiest places in the whole course to pick up evaluation marks, because almost no study fully deserves the jump it makes.
📚 What you need to know
Generalisability is how far findings can be applied to a wider population, setting or time.
Sample-to-population generalisability needs a sample that is random and representative, and large enough.
Inferential generalisability is about applying findings to other settings or groups outside the study.
Theoretical generalisability is about the idea travelling, even when the people cannot.
Qualitative research aims for transferability rather than statistical generalisation.
Generalisability is part of external validity, not a separate idea bolted on beside it.
A big sample helps, but a big biased sample generalises no better than a small one.
The jump every study makes
Researchers study a sample and want to say something about a population. That only works if the sample is a miniature version of the population — same mix of ages, backgrounds, genders, whatever matters for the topic. The moment the sample is lopsided, the conclusion is about a group that does not exist.
Kyle et al. surveyed over thirteen thousand people, which sounds unarguable — until you notice they were all nurses, and all in Scotland.
This is the point students miss most often. Size and representativeness are two different things. Thirteen thousand nurses still tell you about nurses.
Three kinds of generalising
Type
What is being generalised
What it needs to work
Sample to population
The numerical result, from the sample to the wider group
A random, representative and reasonably large sample
Inferential
The finding, to other settings or populations
High external validity, ideally a real-world setting
Theoretical
The underlying idea or explanation, not the numbers
A clear mechanism that other research can build on
Theoretical generalisability is the one worth learning properly, because it rescues research that could never generalise statistically. You cannot generalise from HM as a person — there was only ever one of him. But the principle his case revealed, that damage to a specific brain structure destroys a specific ability, informed decades of memory research. The idea travelled even though the sample could not.
Quantitative and qualitative aims
Quantitative research aims for generalisation.
Qualitative research aims for transferability.
Why it sits inside external validity
Generalisability is not a separate concept sitting alongside validity — it is a component of external validity. External validity asks whether findings apply beyond the research setting; generalisability is the part of that question concerned with who and where. That is why an evaluation point about a narrow sample is really a point about external validity, and saying so out loud is what turns a description into an evaluation.
A useful move in essays: instead of saying “the sample was too small”, say who is missing and why it matters. “The sample was all male university students, so the conclusion may not extend to older adults or to women” is a far stronger sentence, and it is the same length.
🧩 Judging generalisability in four questions
Who was actually studied? Age, gender, culture, occupation, health status.
Who is the conclusion about? Compare that with the answer above — the gap is your evaluation point.
How were they recruited? Volunteers and opportunity samples bring predictable biases with them.
Does the idea still travel even if the people do not? If yes, say so: that is theoretical generalisability.
Worked examples
WORKED EXAMPLE
Evaluate the generalisability
A study of stress and coping recruits 200 first-year psychology students at one university by putting a poster in the department. The researchers conclude that “young adults cope with stress by seeking social support”. Evaluate the generalisability of this conclusion.
Step 1: Describe the sample precisely
First-year psychology students at a single university, recruited by poster, so this is a self-selecting sample.
Step 2: Identify the bias that bringsVolunteer bias — people who respond to posters tend to be more sociable and more interested in psychology, which is directly relevant to a study about seeking social support.
Step 3: Compare sample with conclusion
The claim is about young adults in general, but only educated students from one institution were studied.
Step 4: Give the verdict and a fix
Sample-to-population generalisability is weak, lowering external validity. Recruit across several institutions and include non-students, ideally using stratified sampling.
Volunteer bias plus a narrow sample; conclusion overreachesthe sharpest point is when the sample bias links to the topic itself
WORKED EXAMPLE
Defend a study that cannot generalise
A case study of a single patient with a rare memory disorder is criticised because “it cannot be generalised”. Explain how the research could still be valuable.
Step 1: Accept the criticism honestlySample-to-population generalisation is impossible from one person.
Step 2: Introduce the alternativeTheoretical generalisability: the concepts and explanations developed can inform further research and theory-building.
Step 3: Explain the mechanism
Showing which abilities are lost and which survive reveals how memory systems are organised in everybody, not just in this patient.
Step 4: Add the qualitative standard
The rich description also allows transferability: readers can judge whether the insights apply to similar cases they know.
Theoretical generalisability and transferability, not statistical generalisationthis argument turns a stock criticism into a nuanced evaluation
💡 Exam tip
Name the sampling method when you criticise a sample. It shows you know why the bias is there.
Always link generalisability back to external validity by name.
Use theoretical generalisability whenever a question involves a case study or a very small sample.
Say “transferability” for qualitative research and “generalisation” for quantitative. The vocabulary is assessed.
A good evaluation names the specific group excluded, not just “the sample was unrepresentative”.
Watch out for cultural narrowness — findings from one country are often presented as universal.
⚠ Common mix-up
Thinking a large sample is automatically representative. Size and balance are different properties.
Treating generalisability as separate from validity. It is part of external validity.
Saying case studies have no value because they cannot generalise. They generalise theoretically.
Using transferability for quantitative research. That word belongs to qualitative work.
Forgetting settings and time. Generalising is not only about people; a finding also has to survive a different era.
Writing “unrepresentative” without saying of whom. The detail is where the mark is.
Up next: Bias in the Researcher and the Participant — the two sets of human habits that quietly bend results before anyone has analysed a single number.
Want this explained one-to-one?
Book a free session with an experienced IB Psychology tutor and get your trickiest topics made simple.
