- Why Methodology Matters So Much in a Physics IA
- Physics IA Requirements (Guide First Assessed in 2025)
- Designing a Physics IA Around a Clear Relationship
- Controlling Variables Properly
- Building Uncertainty Into the Method From the Start
- Plan the Linearised Graph Before You Collect Data
- Writing the Conclusion and Evaluation
- How the Four Physics IA Criteria Are Marked
- Physics IA Ideas: Mechanics, Thermal and Electricity
- Physics IA Ideas: Waves, Fields and Nuclear
- Common Physics IA Methodology Errors
Quick Answer
The IB Physics IA is a scientific investigation worth 20% at SL and HL: a report of up to 3,000 words on your own research question, marked on research design, data analysis, conclusion and evaluation, 6 marks each. A strong methodology tests one quantitative relationship, controls variables explicitly, builds uncertainty into the design and plans a linearised graph before any data is collected.
What You'll Learn
- The Physics IA is a scientific investigation worth 20% at SL and HL, with a 3,000-word limit and four criteria worth 6 marks each
- Design around one quantitative relationship that theory predicts, so you can linearise the graph and test it
- Build uncertainty into the method: instrument choice, repeats, propagation, uncertainty bars and maximum and minimum gradient lines
- Explain how each controlled variable is actually controlled, not just list it
- Simple apparatus is fine; the criteria reward rigorous design and analysis, not expensive equipment
Why Methodology Matters So Much in a Physics IA
Your method decides the quality of your data, and the quality of your data limits every other criterion. The research design criterion directly rewards explaining your methodological decisions and describing a method someone else could reproduce. Data analysis, conclusion and evaluation then depend on whether the method produced enough precise data. Many students understand the physics but lose marks because the method is vague, poorly controlled or unable to separate a real trend from measurement uncertainty.
Pro Tip
If your method cannot produce enough good data for a graph with uncertainty bars and a meaningful gradient, the problem is methodological, and no amount of analysis will fix it later.
Physics IA Requirements (Guide First Assessed in 2025)
Physics, biology and chemistry share one IA model in the guides first assessed in May 2025, which apply to everyone sitting exams in November 2026, May 2027 and May 2028, including DP1 students who started in August 2026. The IA is the scientific investigation: an open-ended task in which you gather and analyse data to answer your own research question.
- Weighting: 20% of the final grade at both SL and HL, with the same criteria at both levels
- Length: a report of no more than 3,000 words; charts, diagrams, data tables, equations, calculations, citations, the bibliography and headers do not count
- Time: the guide recommends about 10 hours of teaching time for the IA
- Marking: research design, data analysis, conclusion and evaluation, each out of 6, for a total of 24; marked by your teacher and moderated by the IB
- Approaches: lab experiments, spreadsheet modelling, databases or simulations, alone or combined, as long as you collect and analyse quantitative data
- Group work: optional, in groups of up to three, but each student has their own research question and report, and must not present the same raw data as another student
Watch Out
Guides and videos that mention personal engagement, exploration, analysis and communication criteria, a 6-12 page limit or the group 4 project describe the previous Physics guide. The group 4 project has been replaced by the collaborative sciences project, which is separate from the IA.
Designing a Physics IA Around a Clear Relationship
The strongest Physics IAs test a quantitative relationship that theory predicts, so the analysis can compare your data with a model. For example, a student investigating a pendulum might ask: how does the length of a simple pendulum (0.20-1.00 m) affect its period, and does the relationship match the small-angle prediction for T squared against length? Your method should make clear what changes, what is measured and how everything else is held steady.
- Choose a dependent variable you can measure with realistic precision, using sensors or video analysis where possible
- Use an independent variable range wide enough to reveal the trend, and justify the interval
- State the theory and the assumptions behind it, such as small angles, negligible friction or constant temperature
- Decide in advance which graph should be linear and what its gradient and intercept mean
- Plan enough repeats to estimate random uncertainty at every value
Controlling Variables Properly
Physics experiments often look controlled when they are not. A strong method explains which variables could affect the result, how each is kept constant, and how well. This matters most for timing, friction, alignment, temperature and human reaction time.
- Keep apparatus alignment the same between trials, for example with marked positions or a fixed clamp
- Reduce parallax and reading error by viewing scales at eye level or using digital sensors
- Replace hand timing with light gates or video analysis, or time many oscillations instead of one
- Monitor room temperature or current if they can drift during the experiment
- Say what you could not fully control and how it may affect your conclusion
Watch Out
A table of controlled variables with no explanation of how they were controlled is weak research design.
Building Uncertainty Into the Method From the Start
Top-band physics work treats uncertainty as part of the design, not something added at the end. The physics guide expects you to record uncertainties as a range, propagate them through calculations, show them as uncertainty bars and find the uncertainty in a gradient and intercept.
- 1Find the uncertainty of each instrument, or use half the range of repeated readings if that is larger
- 2Choose apparatus whose precision suits the size of what you measure, so percentage uncertainties stay small
- 3Propagate uncertainties: add absolute uncertainties when adding or subtracting, add fractional or percentage uncertainties when multiplying or dividing, and multiply the percentage uncertainty by the power when raising to a power
- 4Draw uncertainty bars on every processed data point
- 5Draw maximum and minimum gradient lines through the uncertainty bars to find the uncertainty in the gradient and intercept
- 6Use that uncertainty to judge how far your result agrees with theory or an accepted value
Pro Tip
Timing ten oscillations and dividing by ten cuts the percentage uncertainty from reaction time by roughly a factor of ten. Small design choices like this improve data analysis more than extra calculations.
Plan the Linearised Graph Before You Collect Data
Your design should already know what graph you will plot and what its gradient means. Linearising turns a curve into a straight line you can test, and the guide lists linearising graphs and logarithmic scales among the required skills.
- Pendulum: plot T squared against length; the gradient relates to the gravitational field strength
- Falling sphere in oil: plot terminal velocity against radius squared to test Stokes' law
- Intensity and distance: plot intensity against 1 over distance squared to test an inverse-square relationship
- Cooling or radioactive decay: plot the natural log of the quantity against time; the gradient is the decay constant
- Unknown power law: plot log y against log x; the gradient is the power
- Make sure your intervals give evenly spread points on the linearised graph, not just on the raw one
Writing the Conclusion and Evaluation
The conclusion criterion rewards a conclusion that is justified, fully consistent with your processed data including its uncertainties, and compared with accepted scientific context, such as a theoretical prediction or a published value. The evaluation criterion rewards explaining the relative impact of specific weaknesses or limitations and explaining realistic improvements.
- 1State whether the data supports the predicted relationship, using the gradient, intercept and their uncertainties
- 2Compare a derived value with an accepted value and say whether the difference lies within your uncertainty
- 3Identify specific weaknesses, such as timing at the release point or heating of a wire, rather than generic human error
- 4Rank weaknesses by how much they affected the result, and say whether each caused random or systematic error
- 5Give a realistic improvement for each main weakness, such as a light gate or a constant-temperature water bath
- 6Discuss limitations of scope, such as the range tested or assumptions that may fail outside it
How the Four Physics IA Criteria Are Marked
Each criterion is out of 6 and worth 25% of the IA mark. These points paraphrase what the guide describes for the top band of 5-6 marks.
- Research design: the research question is described in a specific and appropriate context, methodological considerations are explained, and the method could be reproduced
- Data analysis: recording and processing are clear and precise, uncertainties are appropriately considered, and processing is appropriate and accurate
- Conclusion: a conclusion relevant to the research question is justified, fully consistent with the analysis, and compared with the accepted scientific context
- Evaluation: the relative impact of specific methodological weaknesses or limitations is explained, with realistic improvements explained for them
Physics IA Ideas: Mechanics, Thermal and Electricity
These ideas suit a strong methodology because each tests a predicted relationship with equipment most schools have. Each is marked SL and HL or HL only according to the theory you need to explain it.
- Large-amplitude pendulum (SL and HL): measure how the period changes with amplitude beyond the small-angle approximation, using a light gate or video analysis.
- Falling spheres in glycerol (SL and HL): measure terminal velocity for ball bearings of different radii and test the Stokes' law prediction that velocity is proportional to radius squared.
- Rubber band launcher (SL and HL): measure launch distance against extension and test whether the stored energy follows Hooke's law or departs from it.
- Rolling objects (HL): compare the acceleration down a slope of cylinders with different mass distributions and relate it to moment of inertia, which is HL-only content.
- Bounce and temperature (SL and HL): measure the ratio of rebound height to drop height for a ball at different temperatures using video analysis.
- Static friction (SL and HL): find the angle at which a block starts to slide as you change the normal force or the contact area, and test whether the coefficient of friction stays constant.
- Cooling and surface area (SL and HL): record the temperature of hot water in containers with different exposed surface areas and compare exponential cooling constants.
- Gas pressure and temperature (SL and HL): heat a sealed flask of air in a water bath, record pressure with a sensor and extrapolate to estimate absolute zero.
- Pencil graphite resistivity (SL and HL): measure resistance against length for pencil lines or leads and compare resistivity across pencil grades.
- Thermistor and temperature (SL and HL): measure resistance over a range of temperatures and linearise by plotting ln R against 1/T.
- Cell internal resistance and temperature (SL and HL): find the emf and internal resistance of a cell from voltage-current graphs at different temperatures.
- Salt solution resistance (SL and HL): measure the resistance of a column of salt solution at different concentrations, keeping temperature and electrode spacing fixed.
Physics IA Ideas: Waves, Fields and Nuclear
Wave and field investigations often produce non-linear relationships, which makes linearisation and uncertainty in the gradient central to the analysis.
- Standing waves on a string (SL and HL): vary the tension and measure the fundamental frequency with a signal generator, then plot frequency against the square root of tension.
- Speed of sound in a resonance tube (SL and HL): measure resonant lengths for different frequencies and find the speed of sound from the gradient. The intercept gives an end correction, which goes beyond the syllabus.
- Refractive index of sugar solutions (SL and HL): measure angles of refraction through a semicircular dish at different concentrations and apply Snell's law.
- Diffraction grating wavelengths (SL and HL): measure the angles of maxima for LEDs of different colours and test the grating equation. Follow your school's laser and LED safety rules.
- Damping of a mass-spring system (SL and HL): attach cards of different areas to an oscillating mass and measure how quickly the amplitude decays.
- Light intensity and distance (SL and HL): measure intensity from a small lamp with a light sensor and test the inverse-square relationship, discussing why a real lamp is not a point source.
- Magnetic field and distance (SL and HL): measure the field from a magnet with a phone magnetometer and find the power law from a log-log graph.
- Induction in a falling magnet (HL): drop a magnet through a coil from different heights and record the peak induced emf with a data logger. Induction is HL-only content.
- Kepler's third law from a database (SL and HL): use published exoplanet data to test the relationship between orbital period and orbital radius for planets of one star type.
- Absorbing beta or gamma radiation (SL and HL): measure count rate through increasing thicknesses of aluminium or lead, correcting for background, only with your school's sources and supervision.
- Water flow from a hole (SL and HL): measure how the horizontal range of a water jet changes with the height of water above the hole, and compare with a prediction from energy conservation.
- Spinning top or fidget spinner (SL and HL): measure how the rotation rate decays over time using video or a phone sensor, and test whether friction gives a linear or exponential decay.
Common Physics IA Methodology Errors
These are the most common reasons a Physics IA stays in the middle bands even when the topic is good. If you want to strengthen your methodology skills, an IB Physics tutor can work through controls, uncertainties and linearisation with you on other experiments; feedback on your own method comes from your teacher.
- Too few values of the independent variable to establish a convincing relationship
- Repeated trials recorded but not used to estimate uncertainty
- Heavy reliance on hand timing without acknowledging or reducing reaction-time error
- Controlled variables named but not genuinely controlled
- No reason given for the chosen range or interval
- A graph that is never linearised, so the relationship is described but never tested
- An evaluation listing problems the method should have anticipated and fixed