How to choose an IA topic you can actually finish (any subject)

To choose an IA topic you can finish, start from something you are genuinely curious about, then put it through four checks: can you get the equipment or data, is there one variable or relationship you can actually measure, does a quick pilot work, and is the scope small enough for the time you have? A topic that passes all four is better than a more impressive one that fails any of them, because a narrow question done thoroughly scores higher than a big one done halfway.

This process works for the sciences, ESS, Maths and Psychology, with the differences set out in a table further down. For subject-specific starting points, see our Biology IA ideas, Chemistry IA ideas and Maths IA topics.

What makes an IA topic one you can finish?

Five properties, in roughly the order you can check them:

Access and scope problems are the usual topic-killers, and both show up in the first week if you look for them.

Step 1: How do you get from an interest to a question?

Start with a list of three to five things you already care about: a sport, a food, a local place, a habit, something from class that surprised you. You will spend weeks on it, and a question you want answered pushes you towards better design decisions.

Then turn each interest into a phenomenon, and the phenomenon into a pair of variables:

• Interest: running. Phenomenon: heart rate takes longer to settle after harder exercise. Variables: exercise intensity and heart-rate recovery time.
• Interest: the river near school. Phenomenon: fewer plants grow near the footpath. Variables: distance from the path and plant species richness.
• Interest: cooking. Phenomenon: vegetables lose vitamins when boiled. Variables: boiling time and vitamin C concentration.

If you cannot name the two variables, you do not yet have a topic, only an area. Keep going until you can. For writing the full question, see how to write an IA research question.

Step 2: What equipment, data or access do you actually have?

Check this before you fall for an idea, not after. Make a short list of what each idea needs and find out whether it exists in your school:

• Lab work: ask the technician or teacher which sensors, colorimeters, balances, water baths and chemicals are available, and whether any reagent is restricted.
• Fieldwork: can you get to the site more than once, do you need permission, and does the season suit it? Our ecology field IA guide covers sites and sampling.
• Databases and secondary data: download a sample of the data now. Many promising data sets turn out to be paywalled, patchy, or too small once filtered.
• Participants: who will take part, and will your school's ethics rules allow it?

An idea that depends on one specific event (a single field trip, a borrowed instrument, a match day) needs a backup plan from the start.

Step 3: Is there one variable you can measure well?

The strongest IAs in the sciences and ESS investigate one relationship: one independent variable changed or selected across a range, one dependent variable measured as precisely as possible, other variables controlled or recorded. In Maths, the equivalent is one clear aim that the mathematics can answer. In the new Psychology IA, it is one focused aim for a study design.

Prefer a continuous measurement to a yes-or-no one. "Germinated or not" throws away information that root length keeps; "in or out" throws away what distance from a target keeps. A continuous dependent variable gives you more to process, more to graph and more to evaluate.

Step 4: How do you pilot a topic?

A pilot is a small trial run of the method at the two extreme levels of your independent variable, done in an hour or two before you commit. It answers three questions:

1. Does the method work at all? Does the end point appear, the sensor log, the survey make sense to participants?
2. Is the effect visible? Is the difference between the extremes clearly bigger than the spread between repeats? If not, widen the range or change the measurement.
3. How long does one run take? Multiply by the number of levels and repeats. If the total does not fit your schedule, cut the design now.

Pilots also give you material for the write-up: the reasons behind your range, your number of repeats and your controls often come straight from what the pilot showed.

Step 5: How narrow should the scope be?

Narrower than feels comfortable. Every extra variable, organism or condition multiplies the work and divides the words you can spend on each. These examples show the same narrowing across subjects:

Which red flags mean you should drop a topic?

• You cannot name the dependent variable or how you would measure it.
• It needs equipment, data or access you have not confirmed.
• It has more than one independent variable "to be thorough".
• The pilot showed no difference between the extremes, or more noise than signal.
• The data take months to collect (slow plant growth, seasonal events) and you have weeks.
• It raises ethics problems: harm or stress to animals, people doing anything beyond everyday activity, sensitive personal data. Ask your teacher before planning further.
• The answer is obvious and there is nothing to process, such as "does more light mean more growth?" with no quantity to calculate.
• You chose it because it sounds impressive, not because you can do it well.
• It overlaps your Extended Essay. The same work cannot count twice.

One red flag can often be fixed by narrowing. Two or more usually means you should move on to your second choice.

How does choosing a topic differ by subject?

The process above applies everywhere, but what the topic has to produce differs:

If you are on the old Psychology course (exams in November 2026), the IA is different; follow your teacher's instructions.

What does a one-week decision plan look like?

At the end of the week you should have a research question with a range, a method you have tested, and a plan for the data. Once you are writing, check your plan against the science IA criteria if you are in Biology, Chemistry or Physics, and if you use AI tools at any point, read how to use AI in the IB honestly first.