IB Physics IA Guide | Experiment Design, Uncertainty Analysis & Top Marks
The IB Physics IA rewards students who can frame a precise, testable question and back it up with rigorous quantitative analysis. Mastering uncertainty propagation and graph linearisation is what separates good reports from great ones.
What Makes an IB Physics IA Score Well—and What Undermines It?
The Internal Assessment(IA) in IB Physics is not simply a lab report. It is a focused scientific investigation where you design the experiment, collect data, analyse results, and evaluate methodology—all under a unified research question. Examiners are looking for genuine scientific thinking, not procedural box-ticking.
The single most reliable predictor of a strong IA is a well-formed research question. Everything downstream—variable control, data collection, uncertainty analysis, and evaluation—depends on it. A vague or untestable question creates structural problems throughout the entire report that even excellent writing cannot rescue.
For a broader overview of how Physics HL fits into your full Diploma, see the IB物理 HL 完全ガイド.
How Do You Choose a Research Question That Is Physically Grounded?
A good Physics IA research question names an independent variable, a dependent variable, and a physical relationship or constant that connects them. It should be specific enough to test experimentally within your school's equipment constraints, and open enough that the outcome is genuinely uncertain.
Weak example: "How does temperature affect resistance?" Stronger example: "How does the temperature of a tungsten filament affect its resistivity, and does the relationship follow a linear model over the range 20 °C to 200 °C?"
The stronger version specifies the material, hints at the theoretical model being tested, and defines a measurable range.
Questions to ask yourself before committing to a topic
| Diagnostic question | Why it matters |
|---|---|
| Can I identify a clear independent variable I can control? | Without control, you are measuring noise |
| Is there a known physical law or model the data should fit? | Allows you to compare experimental and theoretical values |
| Do I have access to the necessary equipment? | Ambitious setups often collapse during data collection |
| Can I collect enough repeated trials to do meaningful statistics? | A minimum of five independent repeats per data point is a rough guide—confirm expectations with your teacher |
| Is the relationship likely to be non-trivial? | Relationships that are "too clean" leave no room to demonstrate evaluation skills |
Topics that tend to generate depth
Physics IA topics with strong analytical potential often involve derived quantities (e.g., resistivity, refractive index, spring constant, coefficient of viscosity) rather than raw measurements. When your analysis requires propagating uncertainties through a multi-step formula, the Exploration and Analysis sections almost write themselves.
How Should You Structure the Exploration Section?
The Exploration section is where you demonstrate scientific understanding before the experiment begins. Examiners are assessing whether you know why your design is appropriate—not just that you followed a procedure.
Independent, dependent, and controlled variables
Name each variable explicitly and justify your choices using physical reasoning.
- Independent variable: What you deliberately change. State the range and the increment, and explain why that range is physically interesting.
- Dependent variable: What you measure. Identify the instrument, its resolution, and its likely sources of uncertainty.
- Controlled variables: Everything else that could affect the dependent variable. For each controlled variable, state how you will control it and why that control matters physically.
Writing a concise hypothesis
Your hypothesis should reference the physical principle you expect to observe. If you anticipate a power-law relationship, say so and cite the underlying theory. If you anticipate a linear relationship, state what the gradient should represent physically. A hypothesis grounded in theory gives your Analysis section a target to compare against.
How Do You Handle Uncertainty Analysis Without Losing Marks?
Uncertainty analysis is the section most students treat as an afterthought and most examiners treat as a window into genuine scientific understanding. Getting it right requires distinguishing between types of error, propagating uncertainties algebraically, and interpreting what the final uncertainty means.
Random vs. systematic errors
| Feature | Random error | Systematic error |
|---|---|---|
| Cause | Statistical fluctuations, reading variability | Calibration offset, zero error, parallax |
| Effect on data | Scatter around the true value | Consistent shift in one direction |
| Detected by | Repeated trials show spread | Comparison with accepted value; all results biased |
| Reduced by | Averaging many trials | Improved calibration; alternative method |
Both types must be discussed in your Evaluation section. Identifying a systematic error but failing to suggest how it could be reduced leaves the evaluation incomplete.
Propagating uncertainties through calculations
When your dependent variable is derived from a formula, you must propagate uncertainties algebraically—not just quote the instrument precision.
For a quantity calculated as a product or quotient, percentage uncertainties add:
If Z = A × B / C, then %ΔZ = %ΔA + %ΔB + %ΔC
For a power relationship, multiply the percentage uncertainty by the exponent:
If Z = Aⁿ, then %ΔZ = n × %ΔA
Show these steps explicitly in your working. An examiner who sees the final uncertainty quoted without derivation cannot award marks for the reasoning behind it.
Using error bars on graphs
Every graph in your IA should include error bars on at least the dependent variable, and on the independent variable if its uncertainty is significant. Error bars are not decorative—they carry information. When you fit a line to your data, draw both the maximum gradient line and the minimum gradient line that can reasonably fit within the error bars. The difference between those gradients is your uncertainty in the gradient itself.
How Do You Linearise a Non-Linear Relationship?
Many physically interesting relationships are non-linear: oscillation period vs. length, decay counts vs. time, drag force vs. velocity. Plotting raw data in a curved graph limits your analysis. Linearisation transforms the relationship so that a straight-line fit yields meaningful physical constants from the gradient and intercept.
Logarithmic linearisation
If you suspect a power-law relationship y = kxⁿ, take the natural log of both sides:
ln(y) = ln(k) + n · ln(x)
Plotting ln(y) against ln(x) gives a straight line with gradient = n and **y-intercept = ln(k)**. You can then extract both the exponent and the proportionality constant from the graph.
Common transformations in Physics IAs
| Relationship type | Raw plot | Linearised plot | Gradient gives |
|---|---|---|---|
| Power law: y = kxⁿ | Curve | ln(y) vs ln(x) | Exponent n |
| Exponential decay: y = Ae^(-kx) | Exponential curve | ln(y) vs x | Decay constant −k |
| Inverse square: y = k/x² | Hyperbola | y vs 1/x² | Proportionality constant k |
| Quadratic: y = kx² | Parabola | y vs x² | Coefficient k |
How Should You Write the Evaluation Section to Demonstrate Real Depth?
The Evaluation is where many IAs plateau. Students list weaknesses without connecting them to the data, or suggest improvements so generic ("use better equipment") that they carry no analytical weight.
A strong Evaluation does three things:
- Identifies the most significant sources of error and estimates their likely impact on the result. Not all errors are equal—prioritise the ones that most affect your conclusion.
- Explains the direction of any systematic bias. If your result is consistently higher or lower than the accepted value, which specific error could account for that bias and why?
- Suggests concrete, physically justified improvements. A specific suggestion—"use a laser gate timer instead of a stopwatch to reduce the reaction-time uncertainty from roughly ±0.2 s to roughly ±0.001 s"—is worth far more than "use more accurate equipment." Note that any numerical values you cite as comparisons should themselves be verified against your actual data and instrument specifications.
Evaluating the model, not just the method
The most sophisticated Evaluations also question whether the theoretical model itself was adequate. Was the assumption of negligible air resistance justified? Was the relationship expected to be linear only within a limited range, and did your data approach that limit? These questions show that you understand the boundary conditions of physical models—a skill that distinguishes genuine scientific thinking.
What Are the Most Common Mistakes That Cost Marks?
Understanding where IAs typically lose marks is as useful as knowing what gains them.
For a comparison of how IA strategies differ across sciences, the IB化学 HL 完全ガイド and IB生物 HL 完全ガイド offer useful parallel perspectives—similar criteria, different analytical tools.
How Does the Physics IA Fit Into Your Broader IB Workload?
The IA is one of the most time-intensive components of the Diploma, but it also runs alongside coursework, examinations, TOK, and CAS. Planning the investigation in stages—research question and background theory first, then pilot experiment, then full data collection, then analysis and write-up—spreads the cognitive load and prevents the late-stage panic that produces rushed Evaluations.
A useful framework: treat the IA as three distinct projects that happen to live in one document. The Exploration is a design project. The Analysis is a mathematical project. The Evaluation is a critical-thinking project. Allocating time to each accordingly prevents the common pattern of thorough data collection followed by a thin, rushed write-up.
For broader time management across all DP components, the IB Diploma の時間管理術 article walks through a practical scheduling approach.
A Note on Changing Requirements and Formats
Exact word counts, mark allocations, assessment criteria language, and submission formats vary by cohort year and school. The guidance in this article reflects enduring principles of experimental physics and scientific writing. Before finalising your IA, always cross-check specific requirements against the official IB Physics Subject Guide for your cohort and confirm expectations directly with your teacher or IA coordinator. What remains stable is the underlying logic: rigorous design, honest data, transparent uncertainty, and evidence-based evaluation.
If you're working through your Physics IA and finding that the uncertainty analysis or linearisation steps need an experienced eye, Quick IB's subject-specialist tutors have gone through this process themselves and can give targeted feedback on exactly those sections.