IB Psychology IA on the new course: how to write the research proposal

On the new IB Psychology course (first exams May 2027), the IA is a research proposal: you design a study of a population you're interested in, using one of the four research methods from your class practicals, and you do not carry it out. It's a written report of up to 2,200 words, marked out of 24 on four criteria worth 6 marks each, and it counts for 30% of your final grade at SL and 20% at HL.

That changes most of what older students remember about the psychology IA. There's no replication, no data collection and no results section. The marks go to how well you justify the study, design it, build the tool you'd collect data with, and think through what the findings could mean, so a modest, realistic idea argued precisely is a strong choice.

What does the new Psychology IA involve?

Each of the four methods is tied to a class practical: interviews to health and well-being, observation to human development, surveys and questionnaires to human relationships, experiments to learning and cognition. You'll have run all four in class before or alongside the IA, which helps: you've already met the practical problems (vague questions, nervous interviewees, observers who disagree) that a strong proposal anticipates.

How is the Psychology IA marked? The four criteria

Below is what each criterion covers and what usually separates a thin answer from a strong one. The official descriptors are in your subject guide; ask your teacher for the criteria sheet and mark your own draft against it.

Criterion A: Introduction (6 marks)

The aim of your study, the real-life problem it addresses, and two relevant studies. What moves you up:

• An aim that names the population, the behaviour and the method. "To investigate phones and sleep" is a topic. An aim says who, what and how.
• A real problem with stakes. Who is affected, and what would someone do differently if they knew the answer?
• Two studies chosen because they shape your design, not because they're famous. For each, say briefly what it found and what that means for your study: it suggests the effect exists, it used a different population, it measured the behaviour in a way you'll improve on.

Criterion B: Research methodology (6 marks)

Your method, your procedure and your ethics. What moves you up:

• A method justified against the alternatives for this aim. Not "questionnaires are quick", but why a questionnaire suits a private behaviour like phone use in bed better than an observation would.
• A procedure someone else could follow: how participants are recruited and sampled, what they're told, what they do, in what order, where and for how long.
• Ethics applied to your population and your procedure, not listed (more on this below).

Criterion C: Data collection (6 marks)

The data-collection tool you design yourself, and the challenges of using it. That means your questionnaire items, your interview schedule, your observation coding scheme, or your experimental materials and test. What moves you up:

• Every part of the tool is explained. Why this item, this wording, this response scale, this behaviour category?
• Clear operationalisation. "Sharing" becomes "child hands a toy to another child without being asked".
• Specific challenges with a response to each: social desirability in self-report on a sensitive behaviour, two observers coding the same act differently, a test so easy everyone scores full marks.

Criterion D: Discussion (6 marks)

Possible findings, their policy implications, researcher bias, and one further method. What moves you up:

• More than one possible outcome. What would it mean if the predicted difference appears, if it doesn't, or if it appears only for some participants?
• Policy implications someone could act on: who would change what.
• Researcher bias that is about you: your expectations, your relationship to the population, the way you'd interpret answers. "I will remain objective" says nothing.
• A further method that answers a question your design can't, with a sentence on what it adds.

A word budget that fits under 2,200

One sensible split (our suggestion, not an IB rule): about 450 words for A, 650 for B, 500 for C and 550 for D, which leaves a margin of around 50. Put the full tool in an appendix if your teacher agrees, but explain its key features in the body, because that's where the marks are.

How do you choose a topic that works?

Because you're not running the study, "works" doesn't mean "I can get 40 participants by Friday". It means you can answer yes to all five of these:

1. Can you name the population precisely? "Teenagers" is too loose. "15–16-year-olds in their first term at a new school" is a population.
2. Is there a real problem someone would act on?
3. Can you find two studies close to your aim? If nothing comes up after a proper search, the idea is too niche. Our guide to finding and evaluating sources for the IA covers where to look.
4. Can you build the tool yourself?
5. Can it be done ethically with that population?

Four ideas that pass, one per method:

Your population doesn't have to be your classmates, since you won't be collecting data from them. But the design must still be realistic and ethical for whoever you choose, and if your idea involves a vulnerable group, talk to your teacher first. Proposals that need brain scans, drugs or deliberately stressing people are very hard to make ethical and realistic; save that material for Paper 1. If you're stuck turning a topic into an aim, the method in how to write a strong IB research question carries over.

What ethics do you need to cover?

Ethics sits in criterion B. A list of principles copied from a textbook earns little; each principle applied to your population and procedure earns a lot. Check the ethics guidance in your subject guide as well. Cover these:

• Informed consent. What participants are told, in what form, and who signs. For participants under 16, many schools and most professional codes require consent from a parent or guardian as well as the young person's own agreement. Find out your school's rule and build it into the procedure.
• Right to withdraw. During the study and afterwards, including having their data removed. Say how: a participant number that only you can link to a name.
• Confidentiality and anonymity. Codes rather than names, where data is stored, who can see it. In a small school, details alone can identify someone ("the only boarder who works weekends"), so say how you'd prevent that.
• Protection from harm. If questions touch on stress, sleep or relationships, what happens if a participant becomes upset? Name the support (the school counsellor, for example) and when you'd stop.
• Deception. Use it only if the study cannot work without it, never about anything that affects whether someone would agree to take part, and always followed by a full debrief. Most proposals need none. Not revealing the exact hypothesis in a memory experiment is a mild case; justify it if you use it.
• Debriefing. What participants are told at the end: the real aim, how their data will be used, how to withdraw it, and who to contact.

Which design choices should you justify?

Every design choice gets three parts: what you chose, why it suits your aim, and what you give up. The choices that matter differ by method.

Experiment

• True or quasi-experiment. If you can randomly allocate participants to conditions (self-testing vs rereading), it's a true experiment. If the groups already exist (bilingual vs monolingual), it's a quasi-experiment, and your causal claims must be weaker.
• Independent samples, repeated measures or matched pairs. Repeated measures removes differences between people but brings order effects, so say how you'd counterbalance. In the vocabulary example, the two word lists must be matched for difficulty.
• Operationalised IV and DV, and controls: standardised instructions, the same time of day, the same test conditions.

Survey or questionnaire

• Closed items (easy to compare, may miss nuance) against open items (rich, harder to code).
• Rating scales with labelled points, and a few reverse-worded items to catch people who agree with everything.
• No leading or double-barrelled items, and anonymous completion to reduce socially desirable answers.

Interview

• Structured, semi-structured or unstructured; one-to-one or focus group.
• How you'd record and transcribe, and how your presence as a fellow student might shape what people say.

Observation

• Overt or covert, participant or non-participant, naturalistic or controlled.
• Time sampling or event sampling, and two observers coding independently so you could check inter-rater reliability.

Do you need statistics in the new IA?

No. You collect no data, so you run no tests. But criterion C is about the tool, and the tool decides what data you would get, while criterion D asks what the findings could be. Being clear about the type of data makes both stronger:

Naming an analysis helps only if it shows you understand your own design; it doesn't replace explaining the tool. Ask your teacher how much detail they expect. At HL, data analysis is assessed directly in Paper 3, covered in our IB Psychology exam guide.

What are the most common Psychology IA mistakes?

• A literature review instead of an introduction. Five studies summarised in turn, none connected to the design. Two, used well, is what the criterion asks for.
• A population of "people" or "students". Everything downstream (sampling, ethics, implications) gets vaguer.
• A method chosen because it's easy rather than because it fits the aim.
• Only the expected finding discussed. A proposal that can't be surprised isn't a research design.
• Running over 2,200 words. The Discussion comes last, so it's what gets lost.
• Letting AI write sections. It's an academic integrity problem and it shows. Read what the IB allows when you use AI in an IA before you open a chatbot.

The others to watch for (a borrowed tool, ethics as a checklist, bias reduced to "I will be objective") are covered under the criteria above. Many are versions of the exam habits in our list of common IB Psychology mistakes, especially generic evaluation and confused methods terms.

A planning timeline

Your school sets the deadlines and roughly 20 hours of class time go to the IA, so treat this eight-week plan as a sequence rather than a schedule.