IB Physics IA Research Questions: Four Tests for the 2025 Syllabus

The IB Physics guide for first assessment 2025 sets the same internal assessment for SL and HL: one scientific investigation, about 10 hours of class time, worth 20% of the final grade. The written report has a maximum of 3,000 words. Charts, data tables, equations and calculations, citations, the bibliography and headers do not count towards that limit.

Four criteria, equal weight.

The report is marked out of 24, with 6 marks each for Research design, Data analysis, Conclusion and Evaluation. Your teacher marks it and the IB moderates the marks. Because the four criteria are equal, a research question that makes the data analysis thin or the evaluation trivial costs you up to half the total before you write a word.

What the guide says about the question.

The research question must be your own, it must involve collecting and analysing quantitative data, and the investigation must have enough extent and depth for every descriptor of the criteria to be addressed. The physics does not need to go beyond the syllabus. A clean investigation of a familiar law, done with care, is a better bet than an ambitious topic you cannot measure.

This article is about choosing and sharpening the question. For the mistakes that cost marks once the investigation is under way, read my IB Physics IA guide.

Most IA ideas start as a topic: pendulums, solar cells, magnets. A topic is not a research question. Before you commit, run your idea through four tests. If it fails one, change the question, not the topic.

Test 1: two variables and a range.

Name the independent variable, its range and its steps, and the dependent variable with the instrument you will use. "How does the release height of a steel ball, from 0.40 m to 2.00 m in steps of 0.20 m, affect its fall time?" passes. "Investigating free fall" does not.

Test 2: theory that predicts the shape.

You should be able to write the equation you expect before you take data. If theory says $h = \frac{1}{2}gt^2$, you know a graph of $h$ against $t^2$ should be a straight line with gradient $g/2$. That single sentence gives you the Conclusion criterion something to compare against.

Test 3: enough data in about 10 hours.

Count the readings. Nine heights with five repeats each is 45 measurements, which is realistic in a lab lesson or two. An idea that needs a week of outdoor readings, or one measurement per day, usually fails here.

Test 4: something real to evaluate.

A good question has a weakness you can quantify: a reaction time, a heat loss, a zero error. If everything works perfectly, the Evaluation criterion has nothing to say.

Make the answer non-obvious.

The guide does not ask you to discover new physics, but "Does $V = IR$ hold for a resistor?" gives a flat conclusion. Asking where a model stops working, or which of two models fits better, keeps the conclusion interesting. Compare "How does current depend on voltage for a resistor?" with "Over what range of current does a filament lamp's resistance stay within 5% of its cold value?"

The syllabus examined from 2025 is organised in five themes: A space, time and motion; B the particulate nature of matter; C wave behaviour; D fields; E nuclear and quantum physics. Good questions exist in every theme. The examples below come from the IB Physics IA ideas library on this site, where each idea has been reviewed and comes with variables, controls and a data source.

A: Space, time and motion.

Kinematics is the easiest place to get clean data, which is also why it is crowded. Pick a question with a twist: "How does the release height of a steel ball, from 0.40 m to 2.00 m, affect the value of g calculated from its fall time?" turns a standard experiment into a test of a systematic error.

If projectiles or uniformly accelerated motion feel shaky, revise them first with the kinematics and projectile motion notes and test yourself on the kinematics worksheet. More ideas are in the kinematics IA ideas list.

B: The particulate nature of matter.

Thermal and gas experiments give strong Evaluation sections because heat loss is real and measurable. "What value of absolute zero follows from extrapolating the volume of a fixed mass of air in a syringe with a free plunger, heated from 20 °C to 80 °C in 10 °C steps at constant pressure?" has a clear accepted value, -273 °C, to compare with.

C: Wave behaviour.

Optics and sound give precise data. "How does the refractive index of a glass prism vary across five wavelengths from about 450 nm to 650 nm, and does a Cauchy relation $n = A + B/\lambda^2$ fit the results?" tests a model, not just a number.

D: Fields.

Database questions are especially good here. "Does $T^2$ against $r^3$ for the four Galilean moons follow a straight line, and what mass of Jupiter results?" uses published orbital data and applies Kepler's third law directly. Review the theory with the gravitational fields and orbits notes and the gravitational fields worksheet, then browse the fields IA ideas.

E: Nuclear and quantum physics.

Radioactivity needs a school source and a GM tube, so check what your lab has. "Does the corrected count rate from a sealed gamma source follow an inverse square law between 3 cm and 30 cm?" is a classic with real background and dead time issues to evaluate.

Yes. The guide lists five approaches that can be used on their own or combined: hands-on laboratory work, fieldwork, a spreadsheet for analysis and modelling, extraction and analysis of data from a database, and a simulation. That matters if your school lab is limited, or if you want a question in astrophysics or particle physics that no school lab could run.

How a database question is judged.

You still need a research question with two variables, and your methodology describes how you selected, filtered and sampled the data instead of how you built apparatus. The Hubble constant and Galilean moons examples above both work this way. Evaluation then focuses on the quality of the data source and your selection, so name the database and justify every exclusion.

Working in a group.

Collaboration is optional, and groups can have at most three students. Each student must still investigate an individual question: a different independent variable, the same independent variable with a different dependent variable, or different data from a shared data set. The report is always individual, and the guide is explicit that no student may present the same set of raw data as another.

Uniqueness.

Each student is expected to formulate, investigate and answer a unique research question. If two classmates want the same idea, change one variable or the system studied, and agree it with your teacher early.

The skills an IA rewards are the same ones Paper 1B tests on an unfamiliar practical. Try this original question, written in the style of the exam, then check it against the markscheme. For more data-based technique, see my guide to IB Physics Paper 1B data based questions.

Question.

A student investigates free fall. A steel ball is released from rest by an electromagnet and falls a height $h$ onto a trapdoor that stops a timer. For nine values of $h$ between 0.40 m and 2.00 m she records the fall time $t$ and plots $h$ against $t^2$. The line of best fit has gradient $4.86 \text{ m s}^{-2}$ with an uncertainty of $\pm 0.10 \text{ m s}^{-2}$, and it crosses the $h$ axis at $h = -0.012 \text{ m}$.

(a) Show that the gradient of the graph is equal to $\frac{g}{2}$. [1]

(b) Determine $g$ with its absolute uncertainty. [2]

(c) Suggest one cause of the non-zero intercept and explain whether it affects the value of $g$ found from the gradient. [3]

(d) A second student's timer adds a fixed delay, so every recorded time is 0.020 s longer than the true fall time. Explain why this error, unlike the one in (c), changes the value of $g$ found from the gradient, and state whether $g$ is too large or too small. [2]

Markscheme

(a)

• $s = ut + \frac{1}{2}at^2$ with $u = 0$ gives $h = \frac{1}{2}gt^2$, so $h$ against $t^2$ is a straight line of gradient $\frac{g}{2}$ ✓

(b)

• $g = 2 \times 4.86 = 9.72 \text{ m s}^{-2}$ ✓
• $\Delta g = 2 \times 0.10 = 0.20 \text{ m s}^{-2}$, so $g = (9.7 \pm 0.2) \text{ m s}^{-2}$ ✓

(c)

• the ball falls slightly further than the measured $h$, for example because the trapdoor must be pushed down a short distance before the timer stops, or because of a zero error in the height measurement, so every true fall is about 0.012 m longer than the recorded $h$ ✓
• this is a systematic error that subtracts the same amount from every value of $h$, shifting the line without changing its slope ✓
• so $g$ from the gradient is unaffected; only the intercept is ✓

(d)

• recorded $t^2 = (t_0 + 0.020)^2 = t_0^2 + 0.040\,t_0 + 0.0004$, so the shift in $t^2$ grows with the fall time and is not the same for every point ✓
• points at larger $h$ move further to the right, the line is less steep and $g$ from the gradient is too small ✓

Why this matters for your IA.

Part (c) is exactly what the Evaluation criterion asks for: a specific limitation, its effect on the result and why the gradient method protects you from it. The answer to part (b) is also consistent with $9.81 \text{ m s}^{-2}$ within its uncertainty, which is the comparison a Conclusion needs.

Parents do not need to understand the physics to see whether the IA is on track. Ask your child three questions in the week the topic is chosen.

1. "What are your two variables, and what range will you use?" A clear answer in one sentence is a good sign. A vague topic means the research question is not finished. 2. "What graph do you expect, and why?" If your child can name the equation behind the graph, the Conclusion criterion has a target.

3. "How many readings is that, and when will you take them?" Compare the answer with the lab lessons the school has scheduled.

Most schools set internal deadlines for the topic, the data collection and the draft, and the teacher gives feedback on one draft. My IB Maths and Physics IA timeline for parents explains what happens at each stage and when outside help is appropriate. The work itself must stay your child's own: a tutor can explain the physics and the criteria, but cannot choose the question or write the report.