IB Physics IA Criteria: How Strong Reports Earn Marks

A Physics IA can contain a clear graph, a sensible experiment and several pages of calculations and still lose marks in the places that matter most.

The problem is usually not a lack of physics. It is the gap between doing an investigation and showing what the investigation allows you to conclude.

Under the current IB Physics course, first assessed in 2025, the scientific investigation is marked out of 24 using four equally weighted criteria: Research Design, Data Analysis, Conclusion and Evaluation. The IA contributes 20% of the final Physics result at both SL and HL.

The report should be no more than 3,000 words. The IB also makes clear that the amount of data should be appropriate to the investigation rather than determined by a universal rule about the number of trials or data points.

That changes how students should approach the write up.

The Current IB Physics IA Criteria

Criterion	Maximum marks
Research Design	6
Data Analysis	6
Conclusion	6
Evaluation	6
Total	24

This is different from the older Physics IA rubric.

Previous guides used Personal Engagement, Exploration, Analysis, Evaluation and Communication. Some otherwise useful online articles, including older Nail IB material, still describe that framework. The current assessment instead uses four criteria worth 6 marks each.

Use older resources for experiment ideas or general physics advice only after checking their marking guidance against the current course.

Research Design: Justify the Experiment You Actually Need

Research Design is not simply the section where you list variables and apparatus.

It assesses whether the methodology is appropriate for answering a focused research question.

A strong design should make the physical relationship clear. The reader should understand what you are changing, what you are measuring, what range you selected and which conditions need to remain controlled.

The procedure also needs enough detail for another reader to reproduce the investigation in principle. The 2026 examiner instructions additionally state that a scientific investigation cannot simply repeat a classic investigation or one listed as part of the course skills, although a prescribed practical can be adapted and extended.

Consider the difference between these approaches.

A weaker design might say that five pendulum lengths were tested and that the period was measured three times.

A stronger design explains why those lengths provide a useful test of the expected physical relationship, why repeated measurements are necessary and which variables could otherwise influence the period.

The improvement comes from justification, not from adding more equipment.

Before collecting data, ask:

• What relationship does physics predict?
• Is my independent variable range capable of testing that relationship?
• Can my apparatus resolve the expected change?
• Which variables could create an alternative explanation?
• Is the amount of data appropriate for the analysis I intend to perform?

The current examiner instructions deliberately avoid prescribing a fixed quantity of data. The amount should suit the investigation.

Personal Engagement Has Changed

You do not need a separate Personal Engagement section under the current Physics IA rubric.

Older advice often recommends explaining why you have always loved basketball, guitars, roller coasters or another object connected to the experiment. That advice came from the previous criterion structure.

There are now no separate Personal Engagement marks.

That does not mean your investigation should feel generic.

Individual thinking can appear through the decisions you make: modifying an apparatus after preliminary observations, selecting a range that tests an interesting part of a model, choosing a more suitable measurement method or adapting a familiar experiment to investigate a different relationship.

Those choices matter when they improve the scientific design.

A paragraph about personal interest does not replace them.

Data Analysis: Uncertainty Cannot Be an Afterthought

Data Analysis is where Physics differs sharply from a report that merely produces a graph.

For the top band, the current rubric expects data recording and processing to be both clear and precise, processing to be appropriate and accurate, and uncertainties to receive appropriate consideration. Correct units, significant figures, decimal places, graph labels and data presentation all contribute to precision.

The Physics teacher support material gives more specific guidance. Students should consider instrument precision, uncertainties, variation, uncertainty bars, maximum and minimum lines where relevant, ranges and the treatment of outliers. It also stresses that not every investigation needs every technique on that list. The treatment should suit the investigation.

This is where many reports become mechanical.

They write an uncertainty beside every instrument in the apparatus table, then never use those uncertainties again.

That is not enough.

A Worked Uncertainty Treatment

The current Physics teacher support material gives the example measurement:

L = (87.4 ± 0.2) cm

The percentage uncertainty is:

percentage uncertainty = (0.2 / 87.4) × 100

percentage uncertainty ≈ 0.23%

The Physics data booklet provides the uncertainty rules used for common calculations. For a quantity raised to a power, the fractional uncertainty is multiplied by the magnitude of that power.

If a processed quantity depended on this length through:

y = L²

then its approximate percentage uncertainty using the booklet rule would be:

percentage uncertainty in y = 2 × 0.23%

percentage uncertainty in y ≈ 0.46%

This is what uncertainty treatment should do: follow the measurement through the processing.

The exact treatment depends on your equation and investigation. If you calculate a gradient, constant, efficiency, resistance or another final quantity, the uncertainty should remain connected to that result rather than disappearing after the raw data table.

The teacher support material also notes that measurement uncertainty should reflect realistic experimental conditions. A digital stopwatch displaying three decimal places does not automatically mean a manually timed event has an uncertainty of 0.001 seconds, because human reaction time may dominate the measurement.

That distinction is important.

Instrument resolution and experimental uncertainty are not always the same thing.

What Separates Mid Band Analysis From Stronger Analysis

A mid band analysis can contain correct calculations while still leaving important gaps.

For example, a student might calculate means, plot a graph and add error bars correctly. But the report may never explain whether the scatter is large compared with the measured trend.

A stronger analysis connects the processing directly to the research question.

It asks whether the data genuinely distinguish between competing possibilities.

Suppose your model predicts a linear relationship. Do not stop after inserting a linear trend line. Examine whether the data support that model within uncertainty. If the gradient has physical significance, interpret it. If a transformation was necessary to test the relationship, explain why you used it.

The goal is not to perform the greatest number of calculations.

It is to perform the calculations that allow you to answer the question.

Conclusion: Use Uncertainty to Decide What You Can Claim

The Conclusion criterion is not a summary of the graph.

The current rubric expects the conclusion to answer the research question consistently with the analysis and to be justified through comparison with accepted scientific context.

This is also where uncertainty becomes interpretive.

Your Data Analysis section may calculate an uncertainty. Your Conclusion should explain what that uncertainty means for the strength of the claim.

Imagine that you determine an experimental value for a physical constant.

Writing:

The experimental value was close to the accepted value.

is weaker than explaining whether the accepted value lies within the uncertainty range of the experimental result.

If it does not, that difference becomes scientifically interesting. It may indicate a systematic effect, a limitation in the model or an experimental condition that has not been adequately controlled.

A strong conclusion therefore contains three connected elements:

• a direct answer to the research question
• evidence from the processed data and its uncertainty
• comparison with accepted physics, supported by a credible source

Do not force the experiment to agree with theory.

An unexpected result can still support strong scientific reasoning if you interpret it honestly.

Evaluation: Stop Writing “Human Error”

Evaluation is not a list of things that went wrong.

The current criterion rewards explanation of the relative impact of specific methodological weaknesses or limitations and realistic improvements connected to those limitations.

That makes statements such as these weak:

Human error affected the results.

More trials would improve accuracy.

Better equipment should be used.

Each statement leaves the important reasoning unfinished.

Which measurement was affected?

Was the effect random or systematic?

Would it increase scatter, shift every result in one direction or change the gradient?

Would more trials actually address that problem?

A stronger evaluation follows a chain:

specific limitation → effect on measurements → effect on final result → realistic improvement

For example, if heat loss is important, do not merely state that some energy escaped.

Explain whether the loss is likely to make the calculated quantity systematically too large or too small. Then propose an improvement that targets that mechanism.

The evaluation should also agree with your analysis. If your data show very little random scatter but a large disagreement with theory, blaming random timing error as the dominant weakness may not fit the evidence.

A Clear Physics IA Structure

The criteria do not require you to use their names as four giant section headings.

A practical report structure can be:

1. Research question and physical context
2. Relevant theory and expected relationship
3. Methodology and variables
4. Raw data and observations
5. Data processing and uncertainty treatment
6. Graphs and analysis
7. Conclusion
8. Evaluation
9. References

Keep only the theory needed to understand the investigation.

Do not spend several hundred words reproducing textbook derivations that never affect the design or interpretation.

The 2026 examiner guidance also confirms that full calculations for every data point are unnecessary. Representative calculations can be sufficient.

Use the saved space for analysis.

Use Old Physics IA Examples Carefully

Older exemplars can still teach you how physicists communicate data, build graphs and discuss experimental limitations.

They cannot automatically show you how the current rubric is applied.

A report submitted before the current course may include a dedicated Personal Engagement section or be organised around Exploration, Analysis, Evaluation and Communication. Those were legitimate requirements under the previous framework.

Current Physics students should instead check every exemplar against Research Design, Data Analysis, Conclusion and Evaluation.

The same caution applies to older online topic lists. Nail IB provides a wide range of possible Physics IA ideas, but its older IA pages still refer to Personal Engagement and the previous criterion structure.

Use the ideas as inspiration, then build your investigation around the current guide.

Where Physics IA Marks Are Really Earned

The four criteria have equal maximum marks, but they are connected.

A poorly designed experiment creates weak data.

Weak data restrict meaningful analysis.

Incomplete analysis limits the conclusion.

An evaluation becomes superficial when it is disconnected from the actual uncertainty and behaviour of the results.

That is why the strongest Physics IAs feel coherent rather than segmented.

The research question leads naturally to the method. The method produces data suited to the question. The analysis shows what those data mean. The conclusion makes only the claims the evidence supports. The evaluation explains what prevented the investigation from doing better.

For broader IA planning, internally link this article to IB Internal Assessment Mistakes Students Must Avoid and the Koncepts IB Biology IA Criteria guide.

For subject practice alongside your investigation, direct students to Koncepts for DP students.