Paper 2: evaluating an unseen study

Discuss an unseen study through two or more concepts for Paper 2 Section B: read it for evidence, choose concepts, anchor every point in its details and argue to a reasoned conclusion.

IB Psychology
Exam skills
SL and HL

20 min readLast reviewed 15 Oct 2026

Section B of Paper 2 gives a short description of a research study you have not seen before and asks you to discuss it with reference to two or more of the six concepts. The study may be experimental or non-experimental and is set in one of the four contexts. The question lists four or more of the concepts, and you choose at least two of them.

The task is evaluation. The examiners already know the study and want to see how well you can use concepts such as bias, causality and measurement to judge how far its conclusions can be trusted. The most important single skill is anchoring: every point must rest on a specific detail of this study, not on what is generally true of studies of this kind.

What the question looks like

A typical question reads: “Discuss the following study with reference to two or more of the following concepts: bias, causality, measurement and/or responsibility.” The study itself is usually a paragraph or two: an aim, a sample, a method and procedure, results with a few numbers, and the conclusion the researchers drew. Ethical details such as consent or deception are often mentioned briefly.

The command term is discuss: a considered and balanced review with a range of arguments, ending in a clear conclusion supported by evidence. In this question the evidence is the study itself.

What the markband looks for	Lower bands	Top band (13–15)
Engagement with the concepts	Basic or some engagement with at least one concept	Very good understanding of and critical engagement with two or more concepts in relation to the study
Analysis	Mostly descriptive; analysis superficial or undeveloped	Well-developed critical analysis
Links to the study	Occasional or sometimes ineffective references to the study	Specific and effective references to the study throughout; links between concepts and the study relevant and well developed
Conclusion	Simplistic or only partly consistent	A reasoned, clearly stated conclusion consistent with the arguments
Terminology	Used, but often imprecisely	Accurate and precise; points relevant, accurate and detailed

Themantic Education (Travis Dixon): a three-step routine for evaluating any study, useful practice for reading an unseen study through its method, sample and conclusions.

A practice study

The study that follows is invented for practice and is used in the examples and the full worked answer. It is set in the learning and cognition context and, like an exam stimulus, contains material for several concepts.

Researchers investigated whether a short daily mindfulness exercise improves adolescents’ working memory. An announcement in a school assembly invited Grade 10 students at one international school to take part, and 62 students volunteered.

The 31 volunteers in two classes whose teacher agreed to run the programme formed the mindfulness group: at the start of every lesson for six weeks, the teacher played a 10-minute recording of a guided breathing exercise. The 31 volunteers in two other classes formed the comparison group and had lessons as normal. Students in the mindfulness classes who had not volunteered read quietly during the exercise.

Before and after the six weeks, every participant completed a digit-span task on a laptop (the longest sequence of digits recalled correctly in order) and rated the statement “I am able to stay focused in lessons” from 1 (never) to 7 (always). The teacher who ran the mindfulness exercise also gave both tests to both groups.

In the mindfulness group, mean digit span rose from 6.1 to 6.8 and mean focus rating from 4.0 to 5.3. In the comparison group, mean digit span rose from 6.2 to 6.3 and mean focus rating from 4.1 to 4.1. No inferential statistics were reported. Parents gave written consent for every participant. The researchers concluded that mindfulness improves working memory and recommended that all secondary schools introduce daily mindfulness exercises.

Reading the study

An unseen study is short, and almost every sentence in it is there for a reason. A stimulus study is usually written to contain material for several concepts: a convenience sample, a self-report measure, a missing control, a strong conclusion, an ethical shortcut. The first job is to find these details before choosing concepts.

Five things to mark

The sample: who, how many, from where, and how they were recruited (volunteers, one school, one country, one gender).
The design: whether anything was manipulated, how participants got into groups (random allocation or not), what was controlled and what was not, whether there was a comparison or control group.
The measures: how each variable was operationalised: a test, a single self-report item, an observation, a physiological measure; who collected the data and whether they knew the hypothesis.
The results and conclusion: the numbers, how big the differences are, whether statistics are reported, and how far the conclusion goes beyond the data (especially words such as “causes”, “proves” or “all”).
The ethics: consent, deception, debriefing, harm, withdrawal, confidentiality, and any vulnerable group.

Details in an unseen study linked to concepts: sample to bias, design to causality, measures to measurement, results and conclusion to change and perspective, consent and harm to responsibility; dashed links from sample to perspective and from results to causality. — Each kind of detail in a study feeds one concept most directly, but some feed several: who was sampled matters for bias and for whose perspective the findings represent, and how big the results are matters for causal claims.

Marking the stimulus with the initial of the concept each detail feeds (B, C, M, R) takes a minute and turns the reading into a plan. Details that feed two concepts are the most valuable, because they let the answer show how concepts interact.

Choosing the concepts

The question lists four or more concepts, and an answer must engage with at least two. Choosing well matters more than covering many. Two or three concepts, each developed with several anchored points, almost always score higher than four or five touched on briefly.

Choose the concepts for which the study gives the most material. A study with volunteers, a non-blind researcher and a self-report measure is rich in bias and measurement; a study with no random allocation and a causal conclusion is rich in causality. Responsibility is a good choice only when the study raises a real ethical issue, not just the presence of consent forms.

Questions to ask of any study

Concept	Questions to ask of the study
Bias	Who is in the sample and who is missing (sampling bias)? Could participants have guessed the aim or wanted to look good (demand characteristics, social desirability bias)? Did the researcher know the conditions or have a stake in the result (researcher bias)? Is there cultural bias in the measures?
Causality	Was the IV manipulated? Were participants randomly allocated? Were confounding variables controlled? If the data are correlational, could the effect run the other way (bidirectional ambiguity) or be due to a third variable? Does the conclusion claim more than the design allows?
Measurement	How was each construct operationalised? Is the measure valid (does it measure the construct) and reliable (would it give the same result again)? Is it self-report, behaviour or physiology? Are the numbers large enough, and are statistics reported?
Change	Does the study measure change over time, and is the period long enough? Do the findings justify the change in policy or practice the researchers recommend? Would the change last?
Perspective	Which approach does the study take, and what does it leave out? Is it etic or emic? Whose viewpoint do the measures reflect? Are the findings generalisable to other cultures?
Responsibility	Were consent, withdrawal, debriefing and confidentiality handled properly? Was anyone harmed or deceived? Is the topic socially sensitive? Is it responsible to recommend a change on this evidence?

Concepts that interact

The top band rewards links between concepts and the study that are well developed, and showing how two concepts interact is one of the clearest ways to develop them. Some combinations come up in almost every study.

Measurement and bias: a self-report measure (measurement) is open to demand characteristics (bias), especially when participants know which group they are in.
Bias and causality: a confounding variable or a non-random allocation is a source of bias that also prevents a causal conclusion.
Causality and change: a recommendation to change policy (change) depends on a causal claim, so a weak design undermines the recommendation.
Measurement and responsibility: researchers have a responsibility not to overstate findings from weak measures, especially when they will be used to change practice.

Anchoring every point in the study

A generic point is one that could be written about almost any study: “the sample was small, so the results cannot be generalised”, “it was a lab experiment, so it lacks ecological validity”. Such points show some understanding of the concept but make no specific reference to the study, and they cap an answer in the middle bands however many of them there are.

An anchored point quotes or paraphrases a specific detail of the study, explains why that detail matters for this study’s aim and conclusion, and considers what limits or offsets the problem. It usually takes three to five sentences.

Five stacked boxes for one evaluation paragraph: point through one concept, evidence quoted from the study (shaded), explanation of why it matters, counterpoint, and link to the study's conclusion. — An anchored paragraph moves from a point made through one concept to a quoted detail of the study, an explanation of why it matters here, a counterpoint, and a link back to the conclusion the researchers drew.

A student writes about the practice study: “The sample was only 62 students, so the results cannot be generalised.” Rewrite this as an anchored point about bias.

Show the solution

Step 1: find the specific detail. The 62 participants were volunteers from one international school, recruited through an assembly announcement.

Step 2: name the kind of bias. This is a self-selected sample, a source of sampling bias. Students who volunteer for a mindfulness study are likely to be interested in mindfulness or to believe it helps.

Step 3: explain why it matters here. Volunteers who expect mindfulness to work may try harder on the post-test digit span and rate their focus higher after six weeks of the exercise, so the improvement may partly reflect their expectations. Students at an international school may also differ from the wider population of “all secondary schools” in the recommendation.

Step 4: counterpoint. Both groups were volunteers, so self-selection affects both equally; it does not by itself explain why only one group improved.

Anchored version. “The 62 participants were volunteers from one international school who responded to an assembly announcement, a self-selected sample. Students who choose to join a mindfulness study may already believe it works, which could raise their effort and their focus ratings after six weeks. Because both groups were volunteers, this cannot fully explain the difference between them, but it does limit generalisation to the ‘all secondary schools’ in the researchers’ recommendation, since students who would not volunteer may respond differently.”

Strengths count too

A discussion is balanced. A study with real strengths, such as a comparison group, pre- and post-tests or an objective measure, should get credit for them, because a judgement on the conclusion depends on weighing strengths against limitations. Strengths must be anchored in exactly the same way: “the digit-span task is an objective, standardised measure of working memory that does not depend on what students think about mindfulness” is anchored; “experiments are scientific” is not.

Structuring the answer

There is no single required structure, but a clear one makes the critical analysis easier to follow and the conclusion easier to justify. The structure that follows works for almost any study and any pair of concepts.

Part	What it does	Length
Introduction	Sum up the study in one sentence (method, sample, main finding, conclusion), name the concepts chosen and state the line of argument	2–3 sentences
Concept 1	Two or three anchored points, including at least one strength or counterpoint	2 paragraphs
Concept 2	Two or three anchored points, ideally showing how it interacts with concept 1	2 paragraphs
Concept 3 (optional)	One or two anchored points, if the study gives strong material	1 paragraph
Conclusion	A judgement on how far the study supports the researchers’ conclusion, drawing the concepts together, and what would make the evidence stronger	3–4 sentences

A full worked example

The full answer uses the practice study on mindfulness and working memory.

Question. Discuss the study with reference to two or more of the following concepts: bias, causality, change, measurement and/or responsibility. [15]

Step 1: read and mark

Detail in the study	Concepts it feeds
62 volunteers from one international school, recruited by assembly announcement	Bias (sampling), change (generalising the recommendation), perspective
Groups formed by class, depending on which teacher agreed to run the programme	Causality (no random allocation; teacher and class confounds), bias
Comparison group with lessons as normal; no placebo activity	Causality, bias (expectations)
Teacher who ran the exercise also gave both tests to both groups	Bias (researcher bias), measurement
Digit span on a laptop; single 1–7 item on focus	Measurement (objective versus single self-report item)
Digit span 6.1 → 6.8 vs 6.2 → 6.3; focus 4.0 → 5.3 vs 4.1 → 4.1; no inferential statistics	Measurement, causality
Non-volunteers in the mindfulness classes read quietly	Responsibility
Conclusion “mindfulness improves working memory”; recommendation to all secondary schools	Causality, change, responsibility

Step 2: choose the concepts and the line of argument

Causality and measurement have the most material, and bias interacts with both. Change is worth a short section because the recommendation to all schools is the study’s boldest claim. The line of argument is that the study gives some evidence of an effect on how focused students feel, weak evidence of an effect on working memory, and too little to support the recommendation.

Step 3: the model answer

Discuss the study with reference to two or more of the following concepts: bias, causality, change, measurement and/or responsibility. [15]

Show a model answer

Introduction. The study compared two classes who did a daily 10-minute breathing exercise for six weeks with two classes who did not, and found larger gains in digit span and self-rated focus in the mindfulness group, leading the researchers to conclude that mindfulness improves working memory. This answer discusses the study through causality, measurement and change, with bias as it affects each, and argues that the evidence supports a modest effect on self-rated focus but not a causal claim about working memory or a recommendation for all schools.

Causality. The study has some features that support a causal conclusion. The IV (the daily exercise) was introduced by the researchers rather than chosen by the students, and there was a comparison group measured before and after, so a general improvement over six weeks, such as students becoming more practised at the digit-span task, should have shown in both groups. The comparison group’s digit span barely changed (6.2 to 6.3), which suggests that practice on the task does not explain the mindfulness group’s rise from 6.1 to 6.8.

However, participants were not randomly allocated: the groups were formed by class, depending on which teacher agreed to run the programme. This makes the study closer to a quasi-experiment, and the teacher is a confounding variable. A teacher willing to start every lesson with mindfulness may also differ in teaching style, enthusiasm or classroom calm, any of which could improve focus and test performance. The comparison group also did nothing in place of the exercise, so the mindfulness group received 10 extra minutes of quiet, structured time and extra attention from their teacher each lesson; the effect could be due to a calm start to lessons rather than to mindfulness itself. Because of these confounds, the study cannot show that the mindfulness exercise caused the improvement.

Measurement. The two outcome measures differ greatly in quality. Digit span is an objective, standardised task that measures one component of working memory, and running it on a laptop means the scoring cannot be influenced by the tester. It is therefore the stronger measure. But the gain of 0.7 digits is small, and with no inferential statistics and only 31 students per group it is impossible to tell whether the difference between groups is larger than chance variation. The conclusion that mindfulness “improves working memory” also generalises from a single span task to working memory as a whole.

The focus measure is much weaker. It is a single self-report item rated from 1 to 7, so its reliability cannot be checked, and it measures how focused students believe they are, not their attention.

This is where measurement and bias interact. The mindfulness group knew they were in the mindfulness condition, and the teacher who ran the exercise gave them the rating scale. Students who had spent six weeks being told the exercise would help them focus may have rated their focus higher to please the teacher or because they expected to improve, a clear case of demand characteristics. The large rise in this group’s focus ratings (4.0 to 5.3), compared with no change at all in the comparison group, is therefore the result most open to bias, even though it is the largest effect.

The same teacher administering both tests to both groups adds a risk of researcher bias, for example in how encouragingly instructions were given. Because the digit-span task was computerised, this risk is lower for working memory than for the focus rating, which strengthens the case that at least part of the digit-span difference is real.

Change. The researchers recommend that all secondary schools introduce daily mindfulness. This is a claim about change in educational practice, and it goes far beyond the data. The sample was 62 volunteers from one international school, students who chose to take part in a mindfulness study and may have been more positive about it than the wider population; students who would not volunteer, or students in schools with different cultures and timetables, might respond differently or not at all.

Six weeks is also too short to show whether any benefit lasts or fades once the novelty wears off. Recommending a change for all schools on this basis raises a question of responsibility, because schools would give up 10 minutes of every lesson on evidence that has not been replicated.

Conclusion. The study gives limited support for its conclusion. Its pre-test, post-test and comparison group design and its objective digit-span measure suggest that something about the mindfulness classes was associated with a small gain in digit span, but the lack of random allocation, the teacher confound and the absence of an active control mean that the cause cannot be attributed to mindfulness itself. The strongest-looking result, self-rated focus, is the most open to demand characteristics.

A stronger test would randomly allocate classes, give the comparison group a matched quiet activity, use a tester who does not know the conditions and report inferential statistics; until then, the recommendation to all secondary schools is not justified.

Why the model answer reaches the top band

Two or more concepts with very good engagement: causality, measurement and change are each developed with several points, and bias is used where it interacts with them.
Specific and effective references to the study: every paragraph quotes details (volunteers from one school, groups formed by class, the same teacher testing, 6.1 to 6.8, 4.0 to 5.3, 31 per group, no inferential statistics).
Critical analysis that is developed: strengths are credited (comparison group, objective task), then weighed against limitations; each point explains why the detail matters for the conclusion.
A reasoned conclusion: it separates what the study does support (a possible small effect on digit span) from what it does not (the cause, the focus result, the recommendation) and says what would make the evidence stronger.

A second study for practice

The second invented study is non-experimental and set in the human development context. A plan is given rather than a full answer.

Researchers investigated whether caregivers’ smartphone use in public playgrounds affects young children’s attention-seeking. Over four weekday mornings, two researchers sat on benches in three public playgrounds in one city and observed 40 caregiver–child pairs, with children judged to be aged 2–5. Caregivers were not told they were being observed. For each pair, the researchers recorded for 10 minutes whether the caregiver was looking at a phone, and counted the child’s “bids for attention” (calling the caregiver, pulling at them, or bringing them an object).

Children whose caregivers used a phone for more than half of the 10 minutes made a mean of 7.2 bids, compared with 3.9 bids for children whose caregivers did not. The researchers concluded that caregivers’ phone use causes children to seek attention and that parents should put their phones away in playgrounds.

Discuss the study with reference to two or more of the following concepts: bias, causality, measurement, perspective and/or responsibility. [15] Write a plan.

Show an answer plan

Concepts chosen: causality, measurement and responsibility, with bias where it interacts.

Causality, strength: naturalistic observation of real behaviour in a real setting, so the difference (7.2 versus 3.9 bids) reflects everyday behaviour rather than a laboratory task.
Causality, limitation: phone use was not manipulated; children who demand attention more may lead tired caregivers to turn to their phones (bidirectional ambiguity); third variables such as the child’s age (estimated as 2–5), tiredness or the time of day may affect both.
Measurement: “bids for attention” is operationalised as three behaviours, which makes it observable, but children’s ages were only estimated and a 2-year-old and a 5-year-old bid differently; no inter-rater reliability is reported for the two observers; “more than half of the 10 minutes” turns phone use into two groups and loses information.
Bias: observers could see whether the caregiver was on a phone while counting bids, so observer bias could inflate bids in the phone group; who visits playgrounds on weekday mornings (caregivers who are not at work) is a specific population.
Responsibility: covert observation in a public place where people expect to be seen by strangers is usually acceptable, but children were the focus and no consent was possible, so no identifying information should be recorded; recommending that parents put phones away on correlational evidence risks blaming caregivers unfairly.
Conclusion: the study shows an association in a natural setting, but measurement problems and the correlational design mean the causal conclusion and the advice to parents are not justified; an observation with blind coders, recorded ages and a design that varies phone use would test the claim.

Summary

Section B gives an unseen study and asks for a discussion through two or more of the four or more concepts listed.
The skill rewarded most is anchoring: every point uses a specific detail of the study and explains what it means for the study’s conclusion.
Reading the study for its sample, design, measures, results and conclusion, and ethics shows which concepts have the most material.
Two or three well-developed concepts score higher than many touched on briefly; showing how concepts interact develops the analysis.
Strengths are credited and weighed, so the discussion is balanced.
Generic points (“small sample”, “lacks ecological validity”) cap the mark unless they are tied to this study’s details and claims.
The conclusion is a reasoned judgement on how far the study supports its own conclusion, and what would strengthen it.

Self-check

To do:
I can explain what the Section B markband rewards and why anchoring matters.
To do:
I can read an unseen study and mark its sample, design, measures, results and ethics.
To do:
I can choose the two or three concepts with the most material in a given study.
To do:
I can turn a generic point into an anchored point with a detail, an explanation and a counterpoint.
To do:
I can show how two concepts interact in a study, such as measurement and bias.
To do:
I can structure a full answer and write a reasoned conclusion about the study’s own claim.

Carry on with Psychology in Augment

With a free account, Augment remembers where you stopped in the notes and which questions you have marked.

Start free